Skip to main content

Full text of "Inorganic seminar abstracts"

See other formats








Return this book on or before the 
Latest Date stamped below. 

University of Illinois Library 

L161 — H41 

Digitized by the Internet Archive 

in 2012 with funding from 

University of Illinois Urbana-Champaign 

1958 - 1959 



CYANOSULFANES - L. C. Thompson 1 

Ronald 0. Rags dale 


Richard L. Carlin 

Robert E. Sievers 

James E. Huheey 


THE HYDRATION OF IONS - A. Wallace Cordes 33 

John G. Verkade 


C. H. Travaglini 








C. Y. Fujikawa 





L. C. Thompson September 23* 1958 

I. Introduction 

Among the compounds containing sulfur and the cyanide group, 
cyanogen thiocyanate (S(CN) 2 ), thiocyanogen ((SCN 2 )), sulfur 
cyanate (S(SCN) 2 ), and disulfur di thiocyanate (S 2 (3CN) 2 ) have long 
been known. The workers who first prepared them regarded these 
substances either as pure or mixed pseudohalogens (halogenoids) or 
as thiocyanate derivatives of SC1 2 or S 2 C1 2 . It was only recently, 
however, that Feher suggested that these compounds be considered 
as cyano derivatives of the sulfanes, that is, as eyanosulfanes 
(e.g., S 2 (SCN) 2 becomes S 4 (CN) 2 -dicyanotetrasulfane) (1). 

II. Preoaration of the four previous ly-known cyanosulfanes 

Dicyanomonosulf ane . This compound has been prepared by the 
traction of aqueous KSCN with an ethereal solution of CNBr (2) 



and by the action of SC1 2 on Hg(CN) 2 , ICN on Ag(SCN), and S 2 C1 W 

on AgCN (3). The last reaction actually gives a mixture of S(CN) 2 

and Sa(CN) 2 . S(CN) 2 can be obtained as white crystals melting at 

62°C by the reaction of S 2 (CN) 2 with Hg(CN) 2 (4). 

B. Dicyanodisulfane (Thiocyanogen ) . Early attempts to prepare 
dicyano disulfane led to the preparation of dicyanomonosulfane and 
are the present methods of preparing the latter compound. The 
first successful preparation of S 2 (CN) 2 was accomplished in I9I9 
by Sdderbiick when he treated an ethereal solution of AgSCN with 
iodine (4). Later experiments have shown that the reaction between 
Hg(SCN)2 and Br 2 in ethyl bromide gives better results (l). The 
£ree halogenoid is obtained by crystallization at -70°C. When 
prepared in this way it melts at -2 to -o°C to a yellow oil which 
polymerizes at room temperatures to the insoluble, brick-red 
parathiocyanogen. Other methods of preparation are to electrolyze 
alcoholic thlocyanates, to allow Mn0 2 to react on H3CN in ether, to 
treat lead(IV) acetate with H3CN (3) and to treat an acetic acid- 
acetic anhydride solution of Pb(SCN) 2 with Br 2 (5)- The analogous 
behaviour to the halogens is noted by -the first two procedures. 
Thiocyanogen has also been prepared by the interaction of N0C1 with 
potassium thiocyanate in liquid S0 2 at -30°C . This product, 
melting at a much higher temperature (15~l6°C) than reported above, 
is claimed to be a more nearly pure substance (6) . 

C. Dicyanotrisulf ane . This was shown to be one of the products 
of the reaction between S 2 C1 2 and Hg(CN) 2 (see above). This com- 
pound is also prepared by treating an ethereal solution of S 2 (CN)2 
with dry H 2 S (7) or by the reaction of SC1 2 on an excess of 
Hg(SCN) 2 (1). 

D. Dicyanotetrasulf ane . S 4 (CN) 2 can be prepared by the action of 
S 2 C1 2 on Hg(SCN) 2 . The compound is isolated at acetone-dry ice 
temperatures in the form of white crystals. It melts at -3o°C, 
but is transformed into a glassy mass upon repeated conversion from 
solid to liquid (8) . 

III . Preparation of New Compounds 

In 1958 Peher presented the results of his investigation 
covering not only the preparation of the compounds described above, 
but also four additional compounds, all of which were now regarded 
as "cyanosulfanes." The preparation of the cyanosulfane S (CN) 2 
(n = 1,2,3,^,5,6,7,8) was accomplished (with the exception of 
S(CN) 2 ) using the reaction represented by the equation (l): 

S X 2 + Hg(SCN)a ■> S y+2 (CN) 2 + HgX 2 (X = CI or Br) 

It should be noted that earlier worker used the same method to 
prepare S 3 (CN) 2 and 3 4 (CN) 2 from the only two dichloro compounds of 
sulfur known to them. 

The determining factor In the reaction appears to be the 
tendency of sulfur to form chains. This is substantiated by the 
fact that the reaction represented by the equation: 

S n X 2 + Hg(CN) 2 ■* HgX 2 + S n (CN) 2 

in which no new sulfur-sulfur bonds are formed, can only be ac- 
complished when n = 1 (l). 

These higher cyanosulfanes are isolated either as yellow-white 
solids (S 5 (CN) 2 , S 6 (CN) 2 )or as yellow-green oils (3 7 (CN) 2 , S 8 (CN) 2 ) 

of high viscosity and a large refractive index. They are quite 
soluble In most organic solvents, but are only slightly soluble in 
methanol, in which they decompose. Water also decomposes them. 
As is characteristic of the lower members of the series, the higher 
cyanosulfanes are polymerized to red, amorphous or glassy, in- 
soluble and infusjble products upon heating. However, the higher 
cyanosulfanes can be kept for weeks at temperatures below 0°C 
without appreciable decomposition (l). 

-^ • MP.t-f-T'"-^ nation of St ructure 

The imc; of physico-chemical properti ?s such as density, re- 
fractive Index, viscosity and surface tension, which give Insights 
into the constitutions of other series of compounds (e.g., the 
chlorosulf&nes J s can not be used in this series since the compounds 
are not stable (to decomposition or polymerization) under the 
conditions of the measurements. 

-' R' Spect ra (l) . The Raman spectra of the cyanosulfanes 
show four regions to which well-defined vibrations can be assigned. 

1. The 2125 cm" 1 regio n. The 2125 cm" 1 region corresponds 
to the C~M valence vibration. As in the dinitriles only one C~N 
* r alence vibration Is found. Analogously only one S-H valence vi- 
bration is to be found in the Raman spectra of the sulfanes. 

2. The 670 cm" 1 region . This region corresponds to the C-S 
valence vibrations. In organic sulfides (R-S-R) and their sulfur- 
analogues (R-S -R) there are two frequencies around 660 and 
650-40 cm" 1 . CH3SCN also shows two frequencies. In all the 
cyanosulfanes there is only one frequency. 


3. The 440 cm"" 1 region. S-S valence vibrations appear in this 

region and are known for a large number of compounds (sulfanes, 
etc. ) . 

4. The ^40 cm** 1 region . Experience with the sulfanes, etc., 
shows that_the deformation frequencies of sulfur chains lie in the 
150-250 cm" 1 region. Consequently all of the bands between 130 
and 325 cm"* 1 are assigned to S-S-S or C-S-S bond angle deformations. 

B. Crystal Study . A crystal structure study of S 3 (CK)2 has shown 
that this molecule has an unbranched chain structure with the middle 
sulfur atom located in a crystallographic mirror plane (9). 

Since the same method of preparation is used throughout and 
since the Raman spectra vary in a consistent manner, it is felt 
that these compounds may be regarded as members of an homologous 
series of cyanosulfanes, NC-S -CN. 

V. Chemical Properties 

No detailed study has been made of the chemical properties of 
any of these compounds with the exception of dicyanodisulfane 
( thiocyanogen) ) Two general properties have been reported, however; 
namely, that all are hydrolyzed by water and that all members 
liberate iodine quantitatively from iodides (1). 

A. Dicyan odis ulfane , This compound has been studied extensively, 
with respect "co its properties as a halogenoid. The organic re- 
actions are summarized by Wood (10). 

1. Oxi dizing properties . S 2 (CN) 2 acts as an oxidising agent 
intermediate between bromine and iodine (E° = -0.77*0 . * s such it 
oxidizes Cu(l) to Cu(ll), As0 3 "" 3 to AsO^ 3 ^ H 2 S to free S, metals 
(even as noble as gold) to MeSCN, and S 2 3 2 to S 4 Q ~ 2 . 

-■ Ha J ogenoid properties . Good reviews are given by W~lden 
idrieth (37 and by" Moeller (11). S 2 (CN) a forms (SCN) 3 ", the 

and Auu. 

Iziterhalogen-halogenoid compounds (SCNCl) 3 , (SCNCl) s , SCNBr, SCNI, 
SCNCI3, SCMbr 3 , and I(SCW) 3 and the internal ogenoid NC •SCN(S(CN) 2 ) • 
NOSCN, corresponding to N0C1, is also formed. The compounds of 
S 2 (CN)p and chlorine have recently been reinvestigated and it has 
been shown that monomeric SCKC1 can be prepared (12) - 

3- Othe r pr op erties . S 2 (CN) 2 has been reported to form com- 
pounds with the hydrogen halides, such as (3CN) 2 '2I-ICI (13). Ad- 
dition products with BF 3 and BC1 3 have also been reported (14). 

B. Dic yano trisulfane and dicyanotetrasulf ane . Solutions of either 
of these uo not react with powdered iron to give thiocyanate unless 
water is piesent. This distinguishes them from S 2 (CN) 2 . However, 
liquid S 4 (CK)2 reacts with iron powder without the addition of 
water and also moderately dissolves sulfur (7) • 

VI. Hydrolysis 

The hydrolysis of the cyanosulfanes has been investigated by 

several workers. Early workers represented the hydrolysis of thio- 
cyanogen as that of a halogenoid (3): 

3(SCN) 2 + 4K 2 ■» H2SO4 + HCN + 5H3CN 

However, this does not begin to show the complexity which is now 
known to characterize this hydrolysis. S = , S 2 3 = , S40 8 = , C0 3 = 
and NH 3 have also been found among the hydrolysis products (14). 
The hydrolysis of S 3 (CN) 2 and S 4 (CN) 2 has been investigated but 
the data were interpreted as showing that the compounds were the 
thiocyanate derivatives of S(0H) 2 and S 2 (OH) 2 , rather than the 
thiocyanate derivatives of the sulfanes, H 2 S and H 2 S 2 as had 
previously been proposed (15). 

All the postulations that had been presented prior to Feher's 
work required the formulation of "fancy :T reactions to account for 
the formation of C0 3 = and NH 3 . Feher, however, has now proposed 
that the hydrolysis of the cyanosulfanes be considered as involving 
the hydrolysis of the C=N group first to the amide (-CONH 2 ), then 
hydrolysis to the acid and finally decarboxylation of the ccid 
leaving.: behind the 3uliane (lo). This explains the formation of 
both C0 3 r and NH 3 very nicely. It should be remarked, however, 
that the hydrolysis of the cyanosulfanes h°.s not been thoroughly 
investigated and that the work that has been done has been con- 
cerned mainly with S 2 (CN) 2 . 

Even in the case cf this seemingly simple compound it is 
readily apparent from the large number of ionic species identified 
in the hydrolysate, that the course of the reaction is very complex 
and that no simple equation can be presented for the hydrolytic 
process . 


Feher and K. Weber, Chem. Ber., 01, 642 (1958). 

F. Audrieth, A. W. Browne and C. W. Mason, J. Am. Chem. Soc, 

2799 (1930). 
Walden and L. F. Audrieth, Chem. Rev., 5, 239 (1923). 
S6derba ; ck, Ann., 419, 296 (1919). 
H. Gardner and H. Weinberger, Inorganic Synth eses Vol. I, 

. McGraw-Hill Book Co., New York, 1°39, p. '34. 
Seel and D. Wesemann, Chem. Ber., 86, 1107 (1953). 
Lecher and M. Wittwer, Chem. Ber.,~55, 1481 (1922) . 
Lecher and A. Goebel, Chem. Ber., 55, 1433 (1922). 
Foss, Acta. Chem. Scand., 10, 1% T1956) . 

L. Wood, Organic Reactions , Vol. Ill, John Wiley and Sons, 
Inc., New York, 1946, p. 240. 

11. T. Moeller, Inorganic Chemistry , John Wiley and Sons, New York, 

12. R. G. R. Bacon, et al . , J. Chem. Soc, 1958 , 764, 774, 782. 

13. E. S6derba"ck, Ann.~,~4"65, 184 (1929). 

14. F. Seel and S. Miiller, Chem. Ber., 88, 1747 (1955). 

15. M. Goehring, Chem. Ber., 7£3, 742 (1945). 

16. Personal communication from F. Feher to L. F. Audrieth. 























Ronald 0. Ragsdale September 30, 1958 

I . Introduction 

According to the Lewis concept, bases, electron pair donors, 
react i^ith acids, electron pair acceptors to form addition com- 
pounds. The synthesis of a large number of molecular addition 
compounds of the Group III elements has been reported in the 
chemical literature. Recently enough information has been ob- 
tained to make possible a comparison of the relative acidity and 
bascity of the Lewis acids and bases comprising these addition 
compounds. The strength of these interactions can be evaluated 
by investigating the variation in the strength of the coordinate 
bond that is formed. The strength of the coordinate bond is de- 
fined as the enthalpy change, £ H, accompanying the gas -phase 
dissociation of the complex (l). It is necessary to work in the 
gas phase in order to eliminate the heat of sublimation and vapor- 
ization of solids and liquids. 

II. Methods of Investigating Acid-Base Interactions 

The most valuable information can be obtained by investigating 
the degree of dissociation of the complex in the gas phase. A 
knowledge of how the degree of dissociation varies with tempera- 
ture permits the calculation of the thermodynamic functions, 

A H, ;'^F, and/^S. Thus, a quantitative measure of the strength 
of the coordinate bond can be obtained (2, 3) • 

The gas-phase dissociation technique is limited because it 
can not be used on compounds which are easily or dif f icultly :&jjs- 
sociated. A greater range may be obtained by the use of calorime- 
ters and suitable solvents (4). 

The relative bond energies have also been established by 
displacement reactions. This method gives only qualitative results 
because a displacement reaction depends on differences in free 
energy between products and reactants and both entropy and sol- 
vation effects conceivably determine the results. 

Saturation pressures (e.g., vapor pressure of liquid or solid 
compounds) have been used to give qualitative indication of dative 
bond strength. 

III. Compounds Formed by Group III Acceptor Molecules 
and Ligands of Group V Atoms 

A. Compounds of Boron 

The following relative order of stability has been found for 
compounds of boron with trimethyl amine (5): (CH3)3N:BF 3 > 

(CH 3 )3N:3F 2 CH 3 > (CK 3 ) 3 N:BF(CH 3 ) 2 > (CH 3 )3N:B(CH 3 )3 (Table, 
Series 1). Since fluorine is more electronegative than a methyl 
group, the acidity of BF 3 would be greater than that of B(CH 3 ) 3 
and the above trend can be easily understood. 

The coordinate-bond strength is also affected ^ the nature 
of the groups on the donor atom. (Table, Series 2) . 

The bond strengths of the boron trifluoride adducts of the 
trimethyl derivatives of Group V elements decrease in the order 
N > P > As > Sb. (Table, Series 5). Borane, however, shows an 
order of coordination with the trimethyl derivations of: 
P ^> N^> As y Sb (6). The order P > N is based on a displacement 

Using boron reference acids with small steric requirements, 
it has been shown (7) that trimethylamine is a better base than 
ammonia and that trimethylphosphine is a better base than phos- 
phine (8) . 

The coordinate -bond strength is not always that to be ex- 
pected from predictions based on Paulings (9) electronegativity 
values. Using pyridine and nitrobenzene as reference bases, it 
has been shown (10) that the acceptor power of the boron halides 
follows the order: BF 3 > 3C1 3 "> BBr 3 . The fact that H 3 As:3Cl 3 
(11) and H 3 As:BBr 3 (12) are known, whereas K 3 As:BF 3 does not form 
even below -100°C. (13) is another less rigorous example. 

B. Compounds of Aluminum 

In general the stability of addition compounds formed between 
ligands of Group V elements and aluminum acceptor molecules is so 
great that the complexes can-not be studied in the gas phase. The 
order of coordination (CH 3 ) 3 N > (CH 3 ) 3 P was established by the 
displacement reaction (1): 

(CH 3 ) 3 P:A1(CH 3 ) 3 + (CH 3 ) 3 N T^r(CH 3 ) 3 P + (CH 3 ) 3 N:Al(CH 3 ) 3 . 

C . Compounds of Gallium, Indium and Thallium 

The bond strengths of the addition compounds of trimethyl - 
gallium and-indium decrease in the order: N > ? > As "> Sb . (Table, 
Series 4 and 5) • The corresponding Tl(CH 3 ) 3 compounds can not be 
studied in the gas phase because of their instability. 

The relative coordinate-bonds strength in the trimethylamine 
addition compounds of Group III elements decrease in the order 
Al > Ga > In > B > Tl (Table, Series 6) . The position of boron in 
the order of acceptor power is at least partly due to steric 


IV. Compounds Formed by Group III Acceptor Molecules 
and LIgands of Group VI Atoms 

A . Compounds of Boro n 

Bases of Group VI atoms form less stable adducts than do 
analogous bases of Group V atoms with Group III acceptor mole- 
cules (14). The compound (CH 3 ) 2 0:3(CH 3 ) 3 does not even exist at 
-78.5§C., but the compound (CH 3 ) 2 3:3(CH 3 ) 3 is known (15). With 
BF 3 the order of coordination is (CH 3 ) 2 > (CII 3 ) 2 S > (CH 3 ) 2 Se 
(Table, Series 7). However, with borane, an acid with steric re- 
quirements even lower than those of BF 3 , the order of coordination 
reverts to S > (Table, Series 8). 

The etherates of boron trifluoride (Table, Series 9) show the 
importance of steric effects. Tetrahydrofur an -boron trifluoride 
is considerably more stable than (C 2 ll5) 2 0:BF 3 . 

B. Compounds of Aluminum 

Etherates of aluminum trialkyls are so stable that it is not 
possible to prepare organo -aluminum compounds by the Gri^nsrd re- 
action. The relative bond strengths follow the series ,; S > 
Se > Te (Table, Series 10) . 

C. Compounds of Gallium, Indium, and Thallium 

For the compounds formed by trimethyl -gallium the order of 
coordination is > Se > S = Te (Table, Series 11) (16), The se- 
quence of coordination is unusual because dime thy lselenide forms 
a less dissociated compound with trimethyl -gallium than does di- 
methyl sulfide . The addition compounds formed between Group VI 
ligands and trimethyl -indium or trimethyl -thallium are much less 
stable than their gallium analogs (l). (CH 3 ) 2 0:In(CH 3 ) 3 and 
(CH 3 ) 2 S:In(CH 3 ) 3 are almost completely dissociated in the vapor 
phase. Of the Group VI compounds of trimethyl -thallium, only 
(CH 3 ) 2 S:T1(CH 3 ) 3 has a sharp melting point. 






& H 
kcal/mole Series 





(CH 3 ) 3 N:BF 3 a 

(CH 3 ) 3 N:BF 2 CH 3 23.1 

(CH 3 ) 3 N:BF(CH 3 ) 2 18.3 

(CH 3 ) 3 N:B(CH 3 ) 3 17-62 

H 3 N:B(CH 3 ) 3 13-75 

CH 3 NH 2 :B(CH 3 ) 3 17-64 
(CH 3 ) 2 NH:B(CH 3 ) 3 19-26 

(CH 3 ) 3 N:B(CH 3 ) 3 17-62 

(CH 3 ) 3 N:BF 3 a 

(CH 3 ) 3 P:BF 3 18-9 
(CH 3 ) 3 As:BF 3 

(CH 3 ) 3 Sb:BF 3 b 

(CH 3 ) 3 N:Ga(CH 3 ) 3 21 
(CH 3 ) 3 P:Ga(CH 3 ) 3 18 
(CH 3 ) 3 As:Ga(CH 3 ) 3 10 
(CH 3 ) 3 Sb:Ga(CK 3 ) 3 - 

(CH 3 ) 3 N:In(CH 3 ) 3 19-9 
(CH 3 ) 3 P:In(CH 3 ) 3 17-1 
(CH 3 ) 3 As:In(CH 3 ) 3 c 

6 (CH 3 ) 3 N 
CH 3 ) 3 N 
CH 3 ) 3 N 
CH 3 ) 3 N 
CH 3 ) 3 N 





;CH 3 ) 2 0:Ga(CH 3 ) 3 9-5 

;CH 3 ) 2 S:Ga(CH 3 ) 3 —8 

;CH 3 ) 2 Se:Ga(CH 3 ) 3 10 

;CH 3 ) 2 Te:Ga(CH 3 ) 3 —8 
Too stable to permit study in the gas phase 
Not formed even at -7S°C . 
Too highly dissociated to be measured 

B(CH 3 ) 3 

A1(CH 3 ) 3 

Ga(CH 3 ) 3 

In(CK 3 ) 3 

T1(CH 3 ) 3 

CH 3 ) 2 0:BF 3 
CH 3 ) 2 3:BF 3 
CH 3 )2Se:BF 3 

CH 3 ) 2 0:3H 3 
CH 3 ) 2 S:3H 3 
CH 3 ) 2 Se:BH 3 

CH 3 ) 2 0:BF 3 

C 2 H 5 ) 2 0:BF 3 

i-C 3 H 7 ) 2 0:BF 3 

CK 2 ) 4 0:3F 3 

CH 3 ) 2 0:A1(CH 3 ) 3 
CH 3 ) 2 S:A1(CII 3 ) 3 
CH 3 ) 2 Se:Al(CH 3 ) 3 
CH 3 ) 2 Te:Al(CK 3 ) 3 











Ref erences 

(1) Stone, F.G.A., Chem. Revs., £8, 101 (l?53). 

(2) 3rcwn, K.C., and Gerstein, 1 i. , J, Am. Chem. Sec, 72, 
2923 (1950)- 

(3) Brown, H.C., Taylor, M.D., and Geistein, M. , J. Am. Chem. 
Soc, 66, I4.3I (l9Mj-). 

{I4.) Brown, H.C*, McDaniel, D.H., and Kafiiger, 0., 

Dete rmination of Organic Structure by Fhysic a 1 Methods , 
edited by E.A. Braude and F.C. Nachod, p . '6'3liT Academic 
Press Inc., New York (1956). 

(5) Burg, A.B. and Green, A, A,, J. Am. Chem. Soc, 65, 1636 

(6) Stone, F.G.A, and Burg, A.B., J. Am. Chem, Soc, 76, 3&6 

(7) Brown, H.C., and Bartholomav, H., J. Am. Chem. Sec., 66, 

(8) Sujishi, S., Thesis, Purdue University, 19i|9» 

(9) Pauling, L., The Mature of the Chemical Bond, Cornell 
University Press, Ithaca, Nov/ York ( 19U5 ) • 

(10) Brown, H.C. 'and Holmes, R.R., J. Am. Chem. Soc, ?6 , 
2173 (1956). 

(11) Stieber, A., Compt. rend., 195, 610 (1932). 

(12) Stock, A., Ber., 3h, 9k9 (1901). 

(13) Martin; D.R. and Dial, R*, J. Am. Chem. Soc, 72, 652 

( II4.) Davidson, N. and Brown, H.C., J. Am. Chem. Soc, 6J+, 
316 (I9lf2). 

(15) Graham, W. and Stone, F.G.A. , J. Inorg. Nuclear Chem., 
3, loij. (1956). 

(16) Coates, G.c, J. Chem. Soc, 1951 , 2003 . 

Reactions of Elemental Sulfur 
N. J. Rose October Ik, 1953 

I. Introduction 

In rhombic and monoclinic sulfur the atoms are arranged in 
puckered octagonal rings (1,2,3**0 • Though the structure in the 
region from the melting point to l60°C has not been clearly 
characterized, the ring structure apparently exists in liquid sul- 
fur just above the melting point (5*7*8,9). Similarly the sulfur 
dissolved in inert solvents probably has the S 8 ring structure as 
is indicated by cryoscopic studies showing sulfur to have a molecu- 
lar weight corresponding to S 8 (6). 

The stoichiometry of reactions involving elemental sulfur (in 
the ring structure) is often very straight-forward and simple. 
Until recently, however, the mechanisms by which such reactions 
take place have been overlooked. The reaction path leading to 
rupture of the S 8 ring and the subsequent separation of individual 
sulfur atoms is of special interest. 

II. Nucleophilic Displacement 

The recent literature contains frequent reference to the 
rupture of the ring by a nucleophilic displacement of sulfur on 
sulfur where a Lewis base or an anion is the nucleophile. 

A. Triphenylphosphine-sulfur Reaction 

On the basis of kinetic studies, Bartlett and Meguerian have 
proposed that the process represented by the equation 

p 3 P: -r S 8 (ring) = 6 3 P:S-S-S 5 -S: (transition state) 

represents the rate determining step in the reaction of 6 3 P with 
S 8 in benzene at room temperature (10). In the next step, another 
O3P molecule attacks the beta sulfur (marked with arrow) forming 
Q3PS (sole final product of reaction) and a new polar species with 
only seven sulfur atoms in the chain. This general process con- 
tinues until the sulfur chain no longer exists. Eight <j) 3 PS mole- 
cules, therefore, result from the rupture of one sulfur ring. 

Kinetic studies showed that the reaction is first order with 
respect to each reactant. This finding is consistent with the 
proposed bi-molecular mechanism for the rate determining step. The 
rate of the reaction was found to increase as the solvent was 
changed from hexane, to benzene, to alpha -bromonaphthalene . This 
general trend would be expected for reactions in which transition 
states have polar characteristics compared to the reactants. In- 
stead of pure benzene, acetonitrile, methanol and phenol were used 
with benzene to provide a mixed solvent for the reaction medium. 
The rates of reaction increased in the order acetonitrile < methanoK 
phenol. This increase does not parallel the increase in the mean 


dlelectrlc constant of the solution but does parallel the relative 
hydrogen bonding abilities of the added substances. The opportunity 
for hydrogen bonding to the terminal sulfur atom in the transition 
state is better than on phosphorus of O3P. The transition state is 
therefore favored which fact accounts for an increase in rate. Para- 
substituents on the phenyl group in the phosphine molecule signifi- 
cantly change the rate in the way expected. The relative order of 
reaction rates is tri-p-tolylphosphine > triphenylphosphine > tri-p- 
chlorotriphenylphosphine. Inductive effects account for the relative 
nucleophilic character of these bases. The relative sensitivity of 
the reaction to changes in nucleophilic character in indicated by 
a value of -2.5 for rho in the Hammett equation compared to a value 
of one for rho in the ionization of substituted benzoic acids (20). 

According to Bartlett, the special stability of the S 8 ring 
with respect to other configurations of sulfur atoms quite possibly 
makes it more resistant to nucleophilic displacement than other 
forms (12). The steps following the transition state formation may 
be rapid for this reason. Also the ensuing reactions require no new 
anion solvation which fact should facilitate rapid reaction in non- 
polar solvents. Another contributing feature might be the fact that 
each attack puts the displaced sulfur atom in the highly stable 
conditions, prevailing in (J) 3 PS . 

The rate of reaction is very rapid with hexatomic sulfur in 
benzene or UV irradiated sulfur in benzene (11). The rates could 
not be measured under the conditions of the primary experiment. 
Perhaps the stability of the S 8 ring can again be invoked as an 

B. Cyanide-sulfur Reaction 

The reaction of CN~ with sulfur in methanol at 40° is me- 
chanistically much the same as the preceding case (13) . The product 
here is exclusively SCN" (14). Kinetically the reaction was fol- 
lowed at the 260 rau absorption of sulfur. Here again the reaction 
was found to be first order with respect to each reactant. The 
arguments supporting the nucleophilic mechanism are similar to 
those invoked for the 3 P system. However, water added to the 
methanol slows down the reaction because of the favored solvation 
of CN~ with respect to the transition state. Dissolved uni- 
univalent electrolytes also slow down the reaction. 

C. Similar Reactions 

Other reactions whose first step or whole mechanism is similar 
to <5> 3 P have been proposed. They include the reactions of S0 3 = (15), 
(R0 2 )0P~ (16), P 4 (18), R 3 N (19), S", SH", HSO3" and OH 7 (1?) with 
sulfur. The reaction of RC=CR with sulfur at' l40°C is reported to 
fall into this category ( 9) . However the state of the sulfur at 
this temperature is presently the subject of rather active contro- 
versy and therefore the olefin mechanism seems somewhat speculative. 

.. ,'. . i 


III. Sulfurtri oxide -sulfur Reaction 

Sulfur and liquid S0 3 form the solid polymer (S-S-O) which is 

thermally unstable (21). At room temperature the substance de- 
composes to sulfur and S0 2 . The solvolysis products of 

(S-S-O) are neither dithionite nor polythionates but S03 = , O30 2 = , 

or S0 4 = . The polymer, as well as S0 3 itself, reacts with dioxane 
and pyridine. These reactions are represented by the following 

(S-S-O) + nBase = nBase:S03 + nS (reactive form) 

When the polymer made by dissolving 35 S in SO3 is itself dissolved 
in dioxane, no radioactivity is detected in the dioxane-S0 3 addition 
product (22). On the basis of these experiments it seems that the 
polymer formula is correct and that the S 8 -SO- : ; reaction features 
the formation of an S-S bond which may have originated by an attack 
of a sulfur atom in the S 8 ring through the donation of electrons to 
the sulfur of S0 3 . A possible first step for the reaction is repre- 
sented by the following equation: 

OS + S 8 (ring) - OS:S-S e -S (intermediate) 

The intermediate may degenerate in the same general way as in the 
<j) 3 P reaction except that the atoms on the chain are now functioning 
as electron donors. The ,„ s Q \ units formed in this degeneration 

could then polymerize to form the final product. If this specu- 
lation is correct, the sulfur ring is a nucleophile in this particu- 
lar reaction. Conversely the ring is an electrophile in the other 
reactions mentioned. 

IV. Conclusion 

In reactions of elemental sulfur, the reaction mechanism must 
take place initially through the rupture of the S Q ring. The special 
significance of this step is clearly demonstrated in the case of the 
CN and $ 3 P reactions with sulfur. 



I - • 



1. T. Moeller, Inorganic C hemistry , John Wiley and Sons, Inc., 
New York, 1952, p". 487. 

2 . N . V . Si dgw I ck , The Che mical Elem e nts a nd Their Compou nds . 
Oxford at the Clarendon Press, 1950, p~. 3'J 5 . 

3. B. E. Warren and J. T. Burwell, J. Chem. Phys,, 3, 6 (1935). 

4. K. Ven Kateswarlu, Proc . Ind. Acad, Sci., 12A, 453 (1940). 

5. R. Norri3, Proc. Ind. Acad. Sci., 16a, 287 (1942). 

6. A. W. Pound and J. R. Pound, J. Phys. Chem. 37, 969 (1933). 

7. R. E. Powell and H. Earing, J, Am. Chem. Soc, 65, 648 (1943) 

8. R. F. Bacon and R. Fanelli, J. Am. Chem. Soc, 65, 639 (1943) 

9. L. Bateman, C. G. Moore and M. Porter, J. Chem. Soc, Aug. 
(1958), 2866. 

10. P. D. Bartlett and G. Meguerian, J. Am. Chem. Soc, 78 , 
3710 (1956). 

11. H. W. Aten, Z. Physik, Chem., 38, 321 (1914). 

12. L. Pauling, Proc Natl. Acad. Sci., 35, 495 (I949). 

13. P. D. Bartlett, and R. E. Davis, J. Am. Chem. Soc, 80, 
2513 (1958). 

14. J. K. Bartlett and D. A. Sloog, Anal. Chem., 2S, 1008 (195*0. 

15. 0. Foss, Acta. Chem. Scand., 4, 404 (1950). 

16. 0. Foss, Acta. Chem. Scand., 1, 8 (1947). 

17. M. Schmidt and G. Talsky, Chem. Ber., jCjO, 1673 (1957). 

18. A. R. Pitochelli, Thesis, U. of 111. (1957). 

19. A. Jennen and M. Hens, Compt. Rend., 242, 786 (1956). 

20. K. J. Laidler, Chemical Kinetics , McGraw-Hill Book Co., Inc., 
New York, 1950, p. 143. 

21. R. Appel and M. Goehring, Z. Anorg. Allg. Chem., 265, 312 

22. R. Appel, Naturwissenschaften, 40, 509 (1953). 


Principles and Applications of Raman Spectroscopy 

Richard L. Carlin October 21, 1958 


When visible light passes through a solution, a very small 
portion of this beam is scattered in different directions. This so- 
called Rayleigh scattering consists only of the wavelengths of the 
incident light — there is no change in frequency or -wavelength. Raman, 
however, discovered (1) a series of very weak lines in the scattered 
radiation whose displacement from the primary line was characteristic 
of the molecular species present. Thus, for each molecule a similar 
pattern of lines, the Raman spectrum, is always found, no matter what 
the incident frequency. It is for this reason that Raman scattering 
is of interest to the chemist. 

Experimental Techniques 

In practice, .visible or ultraviolet light is produced by a 
source such as a mercury arc, passed through a filter, and then into 
the sample bulb. Spectra are then observed in the usual manner at 
right angles to the incident radiation. It is necessary that the 
wavelength chosen avoid regions of absorption or photochemical de- 
composition; this has recently necessitated the use of sources such 
as the sodium vapor lamp. 

The Raman spectra of gases and liquids have been observed, but 
the largest amount of work has dealt with liquids and solutions. An 
obvious advantage of this method is that aqueous solutions can be 
used, since the spectrum of water is much simpler than in the infra- 
red. It is a general rule that Raman spectra are simpler to inter- 
pret than infrared spectra, as the former are much less complex. 
For example, overtones and combination frequencies occur only weakly, 
if at all. 

A principal inconvenience of the method is that all potential 
light-scatterers, such as traces of dust, fluorescent materials, 
and the like, must be removed from the sample. 

Molecular Light-Scattering 

The frequency shifts of the Raman lines relative to the primary 
line are found to be equal to frequencies of vibrational or rotation- 
al transitions of scattering molecules. In Rayleigh scattering the 
incident light quantum (energy hV°) collides with a molecule and is 
simply scattered without change of frequency. In the Raman effect, 
the collision induces the molecule to undergo a transition. If, for 
example, the molecule is in the ground state (n = o) and is excited 
to the first vibrational state (n ■ 1), corresponding to the frequency 
V, the light quantum is scattered with correspondingly diminished 
energy .h(V°-V) . If the transition is" to the ground state from- the 
first excited state, then the quantum will be scattered with corre- 
spondingly greater energy. The two lines are thus the Stokes and 
anti -Stokes line for this frequency. 


The Raman effect should not be confused with fluorescence, 
where first absorption and then re-emission of light take place. 
No absorption takes place here, but only scattering. Similarly, 
while the Raman method is often complementary to the infrared ab- 
sorption technique, the two are basically different in their 
selection rules, and should not be confused. 

Some Simple Applications 

Since the Raman spectrum is charactdristic of each type of 
molecule, it is evident that qualitative and quantitative analysis 
are readily performed. 

Mixed Halides: The Raman spectrum of a mixture of CCI4 and 
CBr 4 , or of SiCl 4 and SiBr 4 , consists of a simple superposition of 
the two components (2); mixtures of SnCl 4 and SnBr 4 give Raman 
spectra containing new lines not attributable to either constituent 
(3). The new lines have been shown to be due to the formation of 
mixed halides in the labile equilibrium: 

SnCl 4 + SnBr 4 -^ SnCl 4 n Brn + SnCl Br 4 . Germanium provides 
an interesting case in this~respect . No mixed halides are observed 
for silicon (4) and this reaction is instantaneous for tin, but 
Delwaulle (3) reports that with a mixture of GeCl 4 and GeBr 4 , there 
is a slow formation of mixed halides which can be followed by taking 
successive Raman spectra. 

Ionic Equilibria: Undissociated nitric acid has a different 
spectrum than the nitrate ion, freed on dissociation. By measuring 
the relative amount of each of these species present as a function 
of hydrogen-ion concentration (6), it is found that nitric acid is 
most highly dissociated at 7M. In a similar manner, perchloric acid 
has been found to be the stronger of the two (7)« 

Principles of Molecular Light -Scattering (8-10) 

The application of an applied electric field E will induce a 
dipole moment P in a molecule by the relation P = aE, where a is 
the polarizability of the molecule. Both P and E are vectors; for 
an isotropic molecule, a is a scalar. In general, an induced moment 
need not be in the direction of the applied field, and a is a 
tensor. The components of P and E are: 

P x ■ a xx E x + a xy E y + a xz E z 

P y " a yx E x + a yy E y + V E z (alJ = aJi) 

P=aE +aE T + aE 
z zx x zy y zz z 

Now, a can be represented by the six coefficients of the 

a X 2 + a Y 2 + a Z 2 + 2a XY + 2a YZ + 2a ZX * 1 . 
xx yy zz xy yz zx 


This is the equation of an ellipsoid, and we shall identify the 
polarizability with this ellipsoid. 

If the axes of the ellipsoid are rotated such that the above 
equation reduces to : 

a xx X2 + a yy Ya + "zz 22 = 1 > 

then the a ±± are called the principal values of a. 

Consider a vibrating molecule in which each atom vibrates 
about its equilibrium position with the same phase and frequency, 
each atom reaches its position of maximum displacement at the same 
time, and each atom passes through its equilibrium position at the 
same time. This vibrational mode is called a normal mode and its 
frequency is known as a normal or fundamental frequency. Generally, 
a molecular vibration is complicated, but it can be shown that any 
such vibration is the superposition of a small number, of normal 
vibrations. Each normal vibration has its own frequency; if two 
or three frequencies are the same, the vibrations are degenerate . 

A simple molecule usually possesses some symmetry elements, 
such as a plane of symmetry or an n-fold rotation axis. For each 
element, there is a corresponding symmetry operation — e.g., re- 
flection in the plane or rotation on the axis. The point group 
denotes the total number of symmetry elements of a molecule. For 
the octahedral case, point group 0, , there are twenty-four such 
elements, for example. n 

If a molecule is distorted in a normal vibration, and then a 
symmetry operation characteristic of the undlstorted molecule is 
performed upon it, three modes are found to be important: 

1) Symmetric modes, where the distorted configuration after 
the operation is indentical with the starting configuration. 

2 ) Antisymmetric modes, where the new distorted configuration 
is equal in magnitude to but opposite in sign from the original 
distorted molecule. 

3) Degenerate modes, where the>- distorted configuration 
is now a linear combinaticn : of two or three normal modes. 

If now, during a vibration, in which the polarizability will 
change, we expand (Taylor's series) the polarizability about the 
center we have: 

a = a + ( ^a/£q) q , 

where the zero subscripts refer to the equilibrium configuration. 
Then, P takes the form, 

P = a Q E + ( da/aq) qE. 


Let, now, the incident light vibrate with frequency V°. The 
first term in this expression also vibrates with frequency V° — 
the source of the Rayleigh scattering. 

In the second term, q oscillates as V and E as V°. Therefore, 
the term represents electric moments oscillating with frequencies 
V°+V. This is the source of the Raman scattering, the intensity of 
which is determined principally by ( 6 a/£ q)'x>» This quantity is 
also a tensor, and it shall be defined as the derived polarizability 
tensor , a ' . 

Consider light, now, incident along the y-axis and scattering 
(Rayleigh) at right angles along the x-axis, with the molecule fixed 
at the origin. Letting the incident light have components E and 
E , each will induce dipole moments P , P , and P of the molecule. 
Since an oscillating dipole does not radiate in its own axial di- 
rection, only P and P will contribute to the scattering along the 
x-axis, producing plani -polarized components of intensities i^ and I_, 
parallel, respectively, to the Y- and Z-axes. For an isotropic 
molecule (polarizability ellipsoid a sphere) i ■ o, and the scatter- 
ing is completely polarized. For non-isotropic molecules, i « o, 
and p = i/I is called the degree of depolarization. 

Since Raman scattering depends on a ' in the same way that 
Rayleigh scattering depends on a Q ^ P' for a Raman line is determined 
by the nature of the appropriate derived polarizability tensor. 

Now, define: 

r a a measure of the overall size of the polarizability 

s_ = a measure of the non-spherical shape of the ellipsoid. 

Then, it can be shown that: 

P - 6S 2 

^5r 2 + 7s 2 

Since r 2 can never be zero for Rayleigh scattering, (3 is always less 
than 6/7; if £ = o, the scattering is completely polarized. 

Analogously, for Raman scattering, 

(3' r 6S' 2 

45r' 2 - + 7S' 2 

r 1 and s 1 may both be zero, though, since a D T is not an ellipsoid. 
When both r' and s' vanish, i = I = o, and the Raman line is 

Therefore, the selection rules for the Raman effect are de- 
termined by a ' = ( <3a/t)q) , which is in turn determined by the 
symmetry of the vibrational modes, which are ultimately determined 
by the shape of the molecule being considered. 


Por a particular vibration, antisymmetric with respect to 
some symmetry operation of the undistorted molecule, the performance 
of this symmetry operation causes the vibrational coordinate q to 
become -q. Since r' * o (ellipsoid same size as before) but s' ■ o, 
then = 6/7 for a Raman line that is not othervjise forbidden. 
This illustrates the selection rule that all permitted Raman lines 
of antisymmetric modes are depolarized. 

If, however, the size and orientation of the ellipsoid remain 
unaltered in a vibration of, for example, an antisymmetric mode, 
then this fundamental is forbidden in the Raman effect. This is 
part of the general selection rule that all vibrations antisym- 
metric with respect to a center of symmetry are forbidden in the 
Raman effect. Furthermore, for a molecule possessing a center of 
symmetry, no vibration can be permitted in both the Raman and infra- 
red absorption spectra. 

All Raman lines of degenerate modes are depolarized. Only 
modes which are symmetric can have fi^C^/li these are usually the 
most intense lines. 

Some Applications 

Selection rules have been worked out for various simple mole- 
cules on the basis of these principles, and these are tabulated at 
the end of the abstract. 

1. The structure of 0s0 4 : The following Raman data have been 
determined for 0s0 4 and two related species (11): 

V 2 + V 4 V 3 Vi 

Os0 4 (l) 335 cm." 1 954 cm." 1 965 cm.~ 1 (pol. 
Re0 4 ~(aq.) 332 916 971 (pol. 

W0 4 =(aq.) 324 833 931 (pol. 

This molecule is expected to be tetrahedral (point group T ,) . The 
selection rules allow four Raman lines, one polarized. Three bands 
are found; V2 + V 4 overlap and are unresolvable . Therefore, 0s0 4 
is probably tetrahedral. Aqueous Re0 4 ~ and W0 4 = give three Raman 
lines, one polarized; ) these are the selection rules for octa- 
hedral molecules (point group 0,). Therefore, it has been proposed 
that these ions are coordinated to two water molecules in solution, 
thereby assuming octahedral structure. Since these species should 
resemble isoelectronic 0s0 4 if the structures are the same, the 
number of Raman lines observed, their position and their degree of 
depolarization should be the same. The data shows this to be true, 
and therefore these ions are indeed tetrahedral in solution. These 
results on 0s0 4 have been independently confirmed. (12) . 

2. XY3 molecules are usually planar or pyramidal. CIP3 and 
BrF 3 have been studied Raman spectrospopically (13) to ascertain 
which of these two structures obtains, although a third possibility 
is a T shape. Electrondif fraction results indicate the pyramidal 
structure, but do not rule out the T-structure. The non-zero 
dipole moment rules out a planar (D3,) symmetry. Previous infrared 
and Raman studies had not considered the T (C2 ) structure. Micro- 
wave spectroscopy has demonstrated that these molecules are 


T -shaped, and this has now been verified by the Raman method. 

3. Woodward and coworkers (14-17) have found that species of 
the type MX 4 ~, where M = Ga, In; X = CI", Br", I~, exist in aqueous 
solution. If they are tetrahedral, they should have four Raman 
lines, one polarized. Sample data are: 

GaBr 4 ~ 

In3r 4 ~ 

T13r 4 ~ 

These data check with isoelectronic molecules of known tetrahedral 

SnBr 4 
InBr 4 ~ 

CdBr 4 = 

The gallium dihalides have also been studied (18-19). They 
have been shown to be tetrahedral in the fused state: 

v 2 


Vx V 3 



210(pol.) 273 



197(pol.) 238 



190(pol.) 209 

v 2 

v 4 

Vi v 3 



220(pol.) 279 



l97(pol.) 239 



l66(pol.) 183 

v 2 

v 4 


v 3 

GaCl 2 ( fused) 





GaCl 4 ~ (aq.) 



346 (pol.) 


GaBr 2 ( fused) 





GaBr 4 ~(aq. ) 





This structure persists in the crystalline state (20) and in benzene 
solution (21) . 



Point Gr< 

AB 2 

Linear Sym. 
Bent Sym. 

C0 2 
H 2 






c s w ' 

AB 3 

Sym. Plane 

BF 3 
NF a 


Planar Y 



° 2 v 
C s 


Plane Square 

CH 4 
AuCJ 4" 

ABC 3 

3 -fold Axis 


c 3 

AB 5 

Trigonal Bi- 



PCI 5 

BrF 5 

D3 h 

AB 6 


SF 6 





D 4d 



IF 7 

Ds h 

Selection Rules (22) 

Point Group No. of Lines No. Pol. Line: 

1 1 

3 2 



3 1 

4 2 

6 3 

6 4 

4 l 

3 1 

6 3 

9 3 

6 2 

9 3 

3 1 

7 2 




1. C. V. Raman and K. S. Krishnan, Nature, 121 , 501 (1928). 

2. M. L. Delwaulle and F . Francois, J. chim. phys., 46, 80 (1949). 

3. M. L. Delwaulle, F. Francois, M. B. Delhaye-Buissett, and 
M. Delitaye, J. phys. radium, 15, 206 (1954). 

4. M. L. Delwaulle and F. Francois, Compt. rend., 219 , 335 (1944). 

5. I. R. Rao, Proc. Roy. Soc. (London), A127, 279 (1930). 

6. 0. Redlich and J. Bigeleisen, J. Am. Chem. Soc, 65, 1883 (1943) 

7- <">. Redlich, E. K. Holt, and J. Bigeleisen, J. Am. Chem. Soc, 
66, 13 (1944). 

8. :> Kizushima, "Eandbuch der ?hysik" ; Vol. 26, ed. by S. Fltfgge, 
Springer-Verlag, Berlin, 1953, p. 1?'I. 

9. W. Wept, "Chemical Applications of Spectroscopy", Vol. 9, 
"Technique of Organic Shomistry", ed, V -/A. Weissberger, 
Int e v h : 1 anc e , N e i • ". fori? , N » y . , 1 956 . 

10. L. A. Woodward, Quart. Rev. (London), 10, 185 (1956). 

11. L. A- Woodward and H. L. Roberta, Trans. Faraday Soc, 52 , 
615 (195*>). 

12. F- «T. Hawkins and W. W. Sabol, J. Chem. Phys., 25, 775 (1956). 

13. K. R. Claassen, B. Weinstock, and J. G. Malm, ibid., 28, 285 
(1953J. "" 

14. L. A, Woodward and P. T. Bill, J. Chem, Soc, 1699 (1955). 

15. L, A. Woodward and A. A. Nord, ibid., 2655 (1955). 

16. h z A. Woodward and A. A. Nord, ibid., 3721(1955). 

17. L. A. Woodward and G. H. Singer, ibid., 716 (1958). 

18. L-. \, Woodward, G. Garton, and H. L. Roberts, ibid., 3723 

19. L. A. Woodward, N. N. Greenwood, J. R. Hall, and I. J. Worrall, 
ibid., 15C5 (1958). 

20. G. Garton and H. M. Powell, J. Inorg. Nuclear Chem., 4, 84 

21. R. K. McMullan and J. D. Corbett, J. Am. Chem. Soc, 80, 
476 (1958). """" 

22. G. Herzberg, "Infrared and Raman Spectra", Van Nostrand, 
New York, New York, 1945, p. 1-11. 

R. £. Si ever s 

I * Introduction 

Mechanism of Octahedral 
Substitution in Methanol 

October 28, 1950 

The elucidation of the mechanisms of reactions of octahedral 
complexes has proven to be a very complex problem. -Attempts have 
been made to categorize these reactions as either Snl (substitu- 
tion, nucleophi lie, unimolecular ) or Sn2 (substitution, nucl cophi lie, 
bimolecular ) * But in practice an unambiguous determination of the 
molecuiarity of such reactions is rarely realized./ Bimolecular 
reactions do not always show second order kineticso 2 In the 
Uniting case, Snl and Sn2 reaction paths should produce different 
stereochemical results. However, the stereochemistry of a reaction 
cannot be predicted without making further assumptions and in 
practice one is only able to expect that Sn2 reactions would be 
more s tereospecif ic than Snl reactions. 3 

In predicting whether to expect an Snl or Sn£ mechanism, 

inese v/o uj 

.it.' a e "cne 

several factors rust be considered, 

electronic: s^.rurrure of the metal complex, size and charge of the 

atoms, so i vent character and enercetic considerations* 

in his review, us 
the electronic structure on 

d valence bond cheery to explain 
the kinetics and 
He concluded that the reaction mechanism 

the effect of 
mechanisms of reactions, 
will depend on the availability of stable inner d orbitals. 

Basolo and Pearson^ have applied crystal field theory in 
their explanation of the kinetic behavior of octahedral complex 
reactions., They have calculated the energies of activation for 
the proposed pentacoordinate and heptacoordina te intermediates 
and conclude that, for Co (III) complexes, the energetically 
most feasible route is through the Snl square pyramid intermediate, 

II* Experimental 

Most of the mechanistic studies of complexes have been carried 
out in aqueous solution. Interpretations of the results are 
complicated by format! en of aquo complexes. Direct replacement 
of one ligand by another is therefore uncommon in water. If a 
proper solvent were chosen, it might then be pc ioible to avoid 
the formation of solvated intermediates, and hence to find 
evidence for a direct bimolecular addition mechanism. 

A. Substitution Reactions of Co(en) 2 Cl 2 + in Me than 


Brown ?nd Incolds made a study of the kinetics of substitu- 
tion of the chloro group in cis -Co(en) 2 by N0 3 ~, Cl~, Br", NCS", 
NOv", r r 3 "" and CH 3 0"",* On the b<isis of this and earlier kinetic 
and stereochemical studies they advanced a unifying hypothesis 
which they termed "edge displacement" mechanism. 

The reactions were carried out in methanol and were followed 
H Dolarimetric, spectroscopic, chemical and radiochemical meth^as. 
It was noted that the first four anions (all weakly nucleophi iir) 
react c* +he seme rate and t he last three react at greatly 


differing rates. Since the first four anions all react at the 
same rate and furthermore at a rate independent of the concen- 
tration of the attacking ion, it v/as concluded that these dis- 
placements proceed by an Snl mechanism. 

To support this view further, it v/as observed that the rates 
of reaction were the some as the rate of racemization of optically 
active d- cis -Co( en) 2 CI? » 

I t was found that for the other three anions (CH 3 0~, N 3 ", and 
N0 2 ~) the second order rates stand in a ratio of 30,000:100:1. 
While Brown and Ingold interpreted the kinetic results as indica- 
tive of a simple Sn2 displacement, Basolo 3 pointed cut that the 
methoxide ion data can be interpreted in terms of an Snl CB 
(conjugate base) mechanism with the methoxide ion acting as the 
conjugate base. 

In order to study the effect of a basic anion per s e, Basolo 
and his coworkers 6 made a study using buffered solutions of trans - 
Co(en) 2 Cl 2 in methanol. They found, for unbuffered solutions, 
a high rate of chloride release in the c"se of basic anions such 
as azide, ntirite and acetate. Buffering the solutions greatly 
reduced these reaction rotes. This supported the belief that 
conjugate base mechanisms are possible in methanol. 

Rate studies were also made on the cis -complex. Buffering, 
here as before, reduced the rates for basic aniens but still 
the rates are higher than for non-basic ani ns. Basolo attributes 
this to ion pair formation. The evidence for ion pair formation 
is based on a shift in the near ultraviolet spectrum observed 
for the c is - isomer. This shift dees not exist for the trans 

Basolo3 has interpreted the results of these investigations 
to indicate an Snl CB mechanism for methoxide icn and either 
an Snl IP (ion pair) or Sn2 (or Sn2 IP) reaction wi th other 
basic anions in the case of the cis complex. For the trans 
complex in its reaction with weakly basic anions he believes an 
Snl dissociation mechanism to be operative. 

B. Substitution Reactions of Cr (NH 3 ) 2 (NCS) 4 ", 

Quite recently, Adamson? has made a kinetic study of the 
substitution reactions of Reinecke's salt in a number of non- 
aqueous solvents. He undertook a study of the role of solvent 
and hydrogen bonding effects on the rates of substitution 
reactions. A complex was chosen that would be soluble in a 
variety of non-aqueous solvents and that would be unlikely to 
become involved in ion-pairing effects. 

The rates of solvolysis were determined for Cr (NH3 ) 2 (NCS) 4 ~ 
in K 2 0, CK3OH, C 6 H 5 OH, D 2 and ni trome thane. For solvents of 
the ROH type the apparent first order rate constant was 0.01 
min. _1 at 60°. In D 2 this rate constant was reduced by about 
50% and it was very small in the non-hydrogen bonding solvent 
ni trome thane. No evidence for a direct bimolecular substitution 
was found when anions such as cyanide, azide and iodidei were 
employed in methanol solutions. 


Adamson explained the results of the solvolysis reactions 
in terms of a mechanism designated Sn2 FS (substitution, nucleo- 
philic, bimolecular: front side attack)* This type of mechanism 
is described as a bimolecular displacement process involving 
cooperative hydrogen bonding interaction between the anion and 
the solvent. 

Principal References 

1. F. Basoio, Chem. Revs », 5§, k$9 (1953)» 

2* A. A. Frost and FU G. Pearson, "runetics end Mechanism 11 , 

John Wiley and Sons, New York (1953)* • 
3o- F. Basoio and R. G. Pearson, "Mechanisms of Inorganic 

Reactions"* John Wiley and Sons^ New York (1958). 
k* H. Taube, Chem. Revs ., 50, 69 (1952). 

5, D. D. Brown and C. K. Ingold, J. Chem. Soc ., 1953 , 267l|. 
6» R. G» Pearson, P. M« Henry and F. Basoio, J. Am. Chern. Soc , 

79, 5379, 5382 (1957). 
7» A. W. Adamson, J. Am. Chem. Soc , 80, 3183 (1958). 


James E. Huheey November 3, 1958 

I. Introduction 

The subject of unusual oxidation states has attracted consider- 
able attention in recent years. Reviews have been published by 
Klemm 1 , Scholder 2 , and Kleinberg. 3 ' 4 ' 5,s It is the purpose of this 
seminar to summarize briefly some of the methods useful in obtaining 
transition metals in unusual oxidation states. Since the elements 
of the first long period are better known than their congeners, they 
will be discussed more extensively. 

Examination of a table of the known oxidation states of the 
transition metals of the first long period shows an unexpectedly 
large number of oxidation states . The maximum oxidation state is 
determined by the number of electrons in the n s and (n-l)d orbitals. 
For these elements the maximum known oxidation state is equal to the 
number of unpaired 3d electrons plus two. For the elements from Sc 
to Mn, each additional electron increases the number of electrons 
which may be either given up entirely or shared with an electro- 
negative element such as fluorine or oxygen. Either type of bonding 
results in a positive oxidation unit of one per electron. For the 
elements from Mn to Zn, each additional step across the periodic 
table removes one d orbital from covalent bonding and increases the 
effective nuclear charge by one, thus increasing the electronegativity 
of the metal . 

One should expect an oxidation state of plus two for all of the 
transition metals. This corresponds to the loss of the 4s electrons. 
All of the metals of the first long period do indeed exhibit this 
state with the exception of Sc, for which only the trivalent state 
is known. 

Concerning the lower oxidation states, several of the metals 
form carbonyls and cyanide complexes in which the formal oxidation 
state of the metal is zero or even negative. 

Two limiting factors that often contribute to the instability 
of a given oxidation state are the oxidizing and reducing properties 
of water. Thus in 1 M acid solutions a substance with an oxidation 
potential greater than 0.00 V or a reduction potential greater than 
1.23 V is thermodynamically unstable. 

II. Preparative Methods 

The simplest type of stabilization of unstable oxidation states, 
but one which is rather fortuitous, is insolubility. Insoluble com- 
pounds tend to be stable. Thus Mn0 2 is the only simple Mn(IV) com- 
pound known. Kleinberg 3 has attributed the stability of the hydra- 
zine complex of Cr(Il) to insolubility rather than to the reducing 
properties of hydrazine. 

■ • . • • • r 

- • 

' ' 8 

. i 

.'■•■' - 







■ • ■ • ' ■ lo no.i 

03qxsni; . . ' ■ 

'■■.' ■ 

• : i • 

8.C ' Ui ■ ■' 

.: "..■' : . 

■ - .. • ■ . , ■■■ 


. - . ■.. . ■ '• • • m 

... . ' ' ' • 


... ■ - : ' . 

•:■ . ' _. ■ 



. -■ 

■■ '. * : - ■ ■• ■ ■• : 

.'. ■ .: ■ . ■■ . 


'■ .. -■ 

. . l -■ 






A change of acid concentration may favor one oxidation state 
over another. Manganate(Vl) ion has a reduction potential of / 2.26V., 
However, its oxidation potential is only -O.56 V. Therefore, 
manganate(Vl) can both oxidize water and destroy itself through dis- 
proportionation to manganous and permanganate ions. However, if the 
pH is kept sufficiently high, the manganate(Vl) becomes a poorer 
oxidizing agent and is stabilized. 

Scholder 7 has taken advantage of reversing the tendency towards 
disproportionation. He calls the phenomenon Symproportionation 
(Symproportionierung) . An example consists of mixing strontium 
chromate, chromium(lll) oxide, and strontium hydroxide in the proper 
proportions to form Cr(IV): 

SrCr0 4 + Cr 2 3 + 5Sr(0H) 2 -*■ 3Sr 2 Cr0 4 + 5H 2 

Another method of stabilizing reactive oxidation states is the 
formation of a complex. In general this stabilizes the desired state 
by giving a particularly stable hybrid and often making the effective 
atomic number (e.a.n.) equal to that of the next inert gas. A well 
known example is: 

Co +2 -> Co +3 + e" E° = -1.84 
Co(CN) 6 " 4 ■> Co(CN) 6 ~ 3 + e" E° = +0.83 

It will be seen that the limiting factor for the lower oxidation 
states is the oxidizing property of the hydrogen ion. If we reduce 
the [H ] , according to the Nernst equation, we should favor the 
stability of a reducing agent. If we choose a more basic solvent, 
qualitatively we should achieve the same effect . Or, to look at it 
from another point of view, according to the Usanovich concept, an 
acid and an oxidizing agent are identical; so a more basic solvent 
such as liquid ammonia should favor the formation of low oxidation 
states. This is indeed the case, for solutions of alkali metals 
have been used to reduce metals to the lower oxidation states. 
Examples are the preparation of zero-valent nickel 8 , zero-valent 
palladium 9 , zero-valent platinum 10 and a cyanide containing both 
Mn(0) and Mn(l). 11 

If the compound we are attempting to prepare is a strong oxi- 
dizing agent, a fused salt system is useful. The higher oxides of 
the transition metals are acid anhydrides and are favored by basic- 
conditions; so fused systems containing an excess of an alkali metal 
oxide may be used. Klemm postulates that the oxide is converted by 
an atmosphere of oxygen to the superoxide which in turn oxidizes 
the metal. 1 ' 12 

3K0 2 + PeO -> K 3 Fe0 4 + 3/2 2 

2K0 2 + NiO -» K 2 Ni0 3 + 2 

2K0 2 + 2CuO ■* 2KCu0 2 + 2 

Another method of limited application entails the thermal 
decomposition of extremely high oxidation states to yield somewhat 
lower ones. Scholder 2 ' 7 has prepared Cr(IV), Cr(V), and Fe(IV) by 












• ■. , 

. ■ ' • I ifii 


. . •■••;.■ ■ . -. ■ 

■ ■ ' ' ' ■ ■ ■. 



... . . , : . ■ , . • , 

a J 


! • 


• ■ 


. ' 


decomposition of barium chromates and ferrate, respectively. This 
method is also useful in obtaining the divalent lanthanons. 

The last method to be discussed is of both theoretical and 
practical interest. Selwood 13 found that decomposition of low con- 
centrations of manganous nitrate on ^--alumina led to the formation 
of the isomorphous manganese (III) oxide. However, if manganous 
nitrate is decomposed on a rutile surface, Mn0 2 is formed. Similarly 
nickel (III) oxide forms on alumina, but nickel ( II) oxide forms on 
magnesia. Jezows'ka-Trzebiatowska 14 has found that' the arsenate ion 
favors the formation of the raanganate(V) ion. 


1. W. Kleram, Angew, Chera. 66, 468 (1954). 

2. R. Scholder, Angew. Chem. 66, 46l (1954). 

3. J. Kleinbert, Unfamiliar Oxidation States and Their Stabi- 
lization . University of Kansas Press (195lJ7~ 

4. J. Kleinberg, J. Chem. Educ . 2£, 32 (1950). 

5. J. Kleinberg, J. Chem. Educ. 2£, 324 (1952). 

6. J. Kleinberg, J. Chem. Educ. 33, 73 (1956). 

7. R. Scholder, Angew, Chem. 65, 240 (1953). 

8. J. W. Eastes and W. Burgess, J. Am. Chem. Soc. 64, 1187 (1942). 

9. J. Burbage and W. Fernelius, J. Am. Chem. Soc. §2, 1484 (1943). 

10. V. Christensen, J. Kleinberg and A. Davidson, J. Am. Chem. Soc. 
75, 2495 (1953). 

11. G. W. Watt, R. E. McCarley and J. W. DawS©, J. Am. Chera. Soc. 
7£, 5163 (1957). 

12. K. Wahl and W. Klemm, Z. anorg. u. allgem. Chem. 270, 69 

13. P. Selwood, J. Am. Chem. Soc. 70, 883 (1948). 

14. B. Jezowska-Trzebiatowska, J. Nawojska and M. Wronska, 
Roczniki Chem. 2J5, 405 (1951). 








. . 


• . 


Recent Chemistry of the Halogens 
Darel K. Straub November 11, I958 

I. Introduction 

Modern work on halogen chemistry has been undertaken largely by 
Schmeisser and co-workers in Germany; Engelbrecht and Cady have also 
contributed to this field. These investigators have prepared and 
characterized a variety of unusual oxygen-containing compounds of the 
halogens. Attention recently has also been directed to the reactions 
and properties of iodine pentafluoride. 

II. Fluorine Derivatives of the Halogen Oxy-Acids 

a. Chloryl fluoride. Chloryl fluoride, C10 2 P, was first pre- 
pared by Schmitz and Schumacher by fractional distillation of the 
reaction products of gaseous fluorine on CIO2 in quartz vessels (l). 
The possibility of violent explosions renders the procedure dangerous. 
Introduction of BF3 into the reaction flask stabilizes the reaction, 
probably due to the formation of C10 2 [BF 4 ] (2). For the preparation 
of larger amounts of pure CIO2F, Schmeisser and Ebenho'ch passed 
fluorine into a concentrated solution of C10 2 in CC1 3 F at -78° (2). 
Chloryl fluoride is a colorless gas that Is very sensitive to moisture. 
It slowly attacks glass at room temperature and reacts with BF3, PF 5 , 
AsF 5 , and SbF 5 to give white compounds having the compositions 
C10 2 [BF 4 ], C10 2 [PF 6 ](m.p. -35°), C10 2 [AsF 6 ] (m.p. 50°), and 
C10 2 [SbF 6 ](m.p. 78°), and with S0 3 to give bright orange C10 2 [S0 3 F] 

b. Perchloryl fluoride, C10 3 F. This substance is one of a 
number of products obtained by the reaction of gaseous fluorine on 
solid KCIO3 at temperatures less than -20° (4,6) . It can be produced 
in larger quantities by the electrolysis of an ice-cooled, 10$ so- 
lution of NaC10 3 in anhydrous HF with a potential of 4-7 volts (5)' 
Perchloryl fluoride is a colorless gas, which hydrolyzes only slowly 
in alkaline solution, in strong contrast to most of the other acyl 
fluorides. It is thermally stable up to 500°, supports combustion, 
and has a characteristically sweet odor. By reaction with ammonia, 
the salt NH4(C10 3 NH) is obtained; organic compounds can be fluorinated 
or perchlorated: 

2C10 3 F / Na 2 C(C00R) 2 -* CF 2 (C00R) 2 / 2NaC10 3 
CIO3F / NaCR'(C00R) 2 •* FCR'(C00R) 2 / NaC10 3 
CIO3F / C e H 6 A 1 C1 3 C 6 H 5 C10 3 / HC1 / A1C1 2 F 

Sodium alkoxides yield ethers: 

CIO3F / 2R0Na •+ R0R / NaF / NaC10 4 (7). 

c. Chloryl oxyfluoride, C10 2 0F. Bode and Klesper in 1951 
isolated from the mixture of products resulting from the reaction of 
gaseous fluorine on solid KC10 3 at -40° or lower, a colorless gas 
having the composition CIO3F (8) . Upon hydrolysis, the substance 
yields fluoride and chloride ions and oxygen, in contrast to perchloryl 
fluoride which yields fluoride and perchlorate ions. Thus, chloryl 
oxyfluoride is an isomer of perchloryl fluoride. However, the C10 2 0F 





produced is contaminated with large amounts of chlorine and even after 
fifteen fractional distillations, Engelbrecht could not obtain a pure 
product (9). The purification is particularly dangerous, since vio- 
lent explosions occur easily. When Bode and Klesper repeated their 
experiment, they evidently obtained perchloryl fluoride instead of 
chloryl oxyfluorlde. So far, C10 2 0F has not been obtained pure, and 
its identity has not been firmly established. 

d. Perchloryl oxyfluoride, C10 3 0F. Gaseous fluorine bubbled 
through 60$ perchloric acid in a quartz apparatus yields CIO4P, along 
with SiF 4 , 2 , and 0F 2 (10). Rohrback and Cady could not obtain the 
compound when an all -carbon apparatus was used. Purification by 
fractional distillation is hazardous. Perchloryl fluoride is a color- 
less gas, which has a sharp acid-like odor and explodes upon freezing 
or contact with grease, water, rubber tubiag, or a flame. 

e. Bromyl fluoride, Br0 2 F. Schmeisser and Pammer in 1955 pre- 
pared Br0 2 F by a method analogous to that used for C10 2 F (11). 
Fluorine was slowly passed into a suspension of Br0 2 in perfluoro- 
pentane at -50°, and the light yellow solid which precipitated was 
fractionally distilled. Bromyl fluoride can also be prepared by 
allowing Br0 2 to react with BrF 5 at -60° to -50° (12). Liquid Br0 2 F 
attacks glass rapidly, explodes with water, and decomposes violently 
at 56 to BrF 3 , Br 2 , and 2 . 

f . Iodyl fluoride, I0 2 F. Iodine pentoxide dissolves in boiling 
IF 5 . Upon cooling, white crystals of IOF3 (or rather, I0 2 [IF 6 ]) 
separate. This compound decomposes at 110° to I0 2 F (13). Schmeisser 
and Lang found that iodyl fluoride also resulted when fluorine was 
passed into a solution of I 2 5 in liquid HF (14) . Iodyl fluoride is 
a white crystalline solid that is stable in dry air and non-hygro- 
scopic. It reacts with AsF s (in HF) to yield white crystals of 
I0 2 [AsF 6 ], and with SbF 5 to yield I0 2 [SbF e ]i 

g. Periodyl fluoride, IO3F. Periodyl fluoride results when 
fluorine is passed through a solution of HI0 4 in liquid HF (14). 
The white crystalline material is much more resistant to hydrolysis 
than I0 2 F and can be kept in glass. It decomposes at 90°-100°, 
releasing oxygen. 

III. Chlorine Nitrate 

If C1 2 and N 2 5 are mixed in equimolar ratio at liquid air 
temperature, and the mixture allowed to stand for some hours at -80° 
and then warmed to room temperature, "chlorine nitrate" is produced 
quantitatively (15) . 

C1 2 / N 2 5 + 2C1N0 3 

The compound CINO3 decomposes when mixed with N0 2 at -5° into N 2 5 
and Cl 2 . This suggests the possibility that it is the nitrate of 
hypochlorous acid. The substance reacts with TiCl 4 (at -80°), 
SnCl 4 (-60°), SiCl 4 , and IC1 3 (-30°) to give chlorine and Ti(N0 3 ) 4 
(white crystals, Sn(N0 3 ) 4 (white crystals, sublimable), Si(N0 3 ) 4 , and 
l(N0 3 ) 3 , respectively (16) . The compound l(N0 3 )3 is a yellow powder, 
which decomposes at 0°. When added to a solution of iodine at 
temperatures below 0°, l(N0 3 ) 3 gives IN0 3 . Chlorine nitrate with 
pyridine yields CI py 2 N0 3 , m.p. 108° (15). 

■ . ' ■ . ... : 

■ . 

. i • . • 

q 691 ! .: . . . 

■ ' . "■ 

: ■ '.•.'..';■. 

"'■'■'■-■ '■ .■■■■"'■ 

■ ■ : - . . . 

.'. . .. • . ■ 

' ' ' 


■'■ ' ■ ■ . 

.. ■ I aubq, ' • • • 

■ i ' , J , ■ cisJ 

'■ ■ • 

■ ■ , 

. • •-. 

•• j . 

I ' . 9 . * 

■ ■ • ■ - , ■ 

■ ■.'•- ■ ■ . - 

. ■'• - . • 

■ - 



IV. Fluorine Derivatives of Non-Halogen Acids 

a. "Fluorine fluosulf onate". Fluorine will not react with sul- 
furic acid, but it does react with sulfamic acid above 200° to give 
SO3F2 (17). This compound can be more easily prepared by the reaction 
of fluorine with S0 3 . Infra-red and nuclear magnetic resonance 
studies indicate the structure to be FOSO2F (18). 

Other recently investigated compounds with sulfur and fluorine 
are S0F 4 and S0F 6 , both produced by the fluorination of thionyl 
fluoride (19). The structures F 4 S0 and SF 5 0F are assigned on the 
basis of NMR studies (18). 

b. Trifluoroacetyl hypofluorite, F3CCOOF. Cady and Kellog ob- 
tained the compound F3COOOF when they permitted water vapor, fluorine, 
and F3CCOOH to react (20). The water seemed to act as a catalyst. 
The compound decomposes completely within a few hours at room 
temperature . 

V. Reactions with Iodine Pentafluorlde 

Iodine pentafluoride has attracted some attention recently as a 
non-aqueous solvent. A few of its properties are listed in the table. 
There is evidence for neutralization reactions in IF 5, although no 
pure products have been obtained (21). Both KF and SbF 5 dissolve to 
produce conducting solutions, from which KIF 6 and SbF^IFs] can be 
isolated. When these solutions are mixed and the solvent removed, 
KSbF 6 -0.23IF 5 remains . The compound KIF 6 was the first reported well- 
defined polyhalide containing seven atoms (22). Potassium fluoborate 
can be prepared by passing BF 3 into a solution of KF in IF 5 (21). 
Potassium iodate dissolves readily in cold IF 5 , without decomposition, 
and behaves as a normal salt does In water. It is interesting to note 
that iodine solutions in IF 5 are chocolate brown in color, Indicating 
extensive solvolysis (23). 

Iodine pentafluoride reacts with P2O5, V2O5, Cr 2 3 , W0 3 , and 
KMn0 4 to give P0F 3 , V0F 3 , CR'fceF 2 , W0 3 «2IF 5 , and Mn0 3 F, respectively 
(23,24). Nitrogen (IV) oxide dissolves readily to give a deep cream- 
colored, crystalline solid, IF 5 'N0 2 , which sublimes unchanged when 
gently heated. One of the most interesting reactions is that with 
mercury. Mercury slowly dissolves in refluxing XF 5 to give a dense, 
buff -colored solid, Hg(lF 5 )2, which turns red on exposure to air. It 
is soluble in alcohol, insoluble in chloroform, and hydrolyzes slowly 
in water. With NaOH solution, yellow HgO, fluoride ion, and iodate 
ion are produced (23). 







Table 1. 
Physical Properties of the Halogen Oxy Compounds 









C10 2 F 










C10 2 0F 






-167. 3? 




Br0 2 F 


dec . 56 



I0 2 F 










N0 2 0C1 





FOS0 2 F 

















-21 . 5- 



* Calculated From Cooling Curve 
** At 755 mm Hg 
*** Extrapolated 

Table II. 

Properties of Iodine Pentafluoride (26,27) 

M.P. 9-6° 

B.P. 98 

Density at 0° 3-75 g/ral 

Dielectric constant at 25° 36 

Specific conductance in HF at 25° 1.53*10" 5 / onm cm 




r . 



1. H. Schraitz and H. J. Schumacher, Z. anorg. allgera. Chem., 249 , 
238 (1942). 

2. M. Schmeisser and F. L. Ebenh6*ch, Angew, Chem., 66, 230 (1954). 

3. M. Schmeisser and W. Fink, Angew, Chem., 69, 780~Tl957) • 

4. H. Bode and E. Klesper, Angew. Chem., 66,~o05 (1954). 

5. A. Engelbrecht and H. Atzwanger, Monatsh., 83, 1087 (1952). 

6. J. E. Sicre and H. J. Schumacher, Angew, Chem., 60,, 266 (1957). 

7. Pennsalt Chemicals Corp., New Products , "PF Perchloryl fluoride" 


8. H. Bode and E. Klesper, Z. anorg. allgem. Chem., 266, 275 (1951) 

9. A. Engelbrecht, Angew. Chem., 66, 442 (1954). 

10. G. H. Rohrback and G. H. Cady, J. Amer. Chem. Soc, 6^, 677 

11. M. Schmeisser and E. Paramer, Angew. Chem., 6j_, 156 (1955)* 

12. M. Schmeisser and E. Pamraer, Angew. Chem., §£, 781 (1957). 

13. E. E. Aynsley, R. Nichols, and P. L. Robinson, J. Chem. Soc, 
1953 , 623. 

14. M. Schmeisser and K. Lang, Angew. Chem., 67, 156 (1955). 

15. M. Schmeisser, W. Fink, and K. BrSndle, Angew. Chem., 6^, 780 

16. M. Schmeisser and K. BraVidle, Angew. Chem., 6£, 781 (1957). 

17. F. B. Dudley, G. H. Cady, and D. F. Eggers, Jr., J. Am. Chem. 
Soc, 78, 290 (1956). 

18. F. B. Dudley, J. N. Shoolery, and G. H. Cady, J. Am. Chem. Soc, 
78, 568 (1956). 

19. F. B. Dudley, G. H. Cady, and D. F. Eggers, Jr., J. Am. Chem. 
Soc, 78, 1553 (1956). 

20. G. H. Cady and K. B. Kellog, J. Am. Chem. Soc, 75, 2501 (1953). 

21. A. A. Woolf, J. Chem. Soc, 1950 , 3678. 

22. H. J. Emeleus and A. G. Sharpe, J. Chem. Soc, 1949 , 2206. 

23. E. E. Aynsley, R. Nichols, and P. L. Robinson, J. Chem. Soc, 
1953 , 623. 

24. E. E. Aynsley, J. Chem. Soc, 1958 , 2425. 

25. H. Martin and Th. Jacobsen, Angew. Chem., 6>£, 524 (1955). 

26. M. T. Rogers, H. B. Thompson, and J. L. Spiers, J. Am. Chem. 
Soc, J6, 4841 (1954). 

27. R. C. Brasted, "The Interhalogens", p. 210 in M. C. Sneed, 
J. L. Maynard, and R. C. Brasted, (Ed.), Comprehensive 
Inorganic Chemistry , vol. Ill, D. Van Nostrand Co., N. Y., 1954. 


■ ■ ' 






' . . 










' ■ . . ■ ,■'■.■■ 

. . . ■ ' 

• • •■ . 

"• . " 



. •- 






. ' 




• ■ • 

■ ■ 



A. Wallace Cordes November 18, 1958 

I. Introduction 

It has always been assumed that ions in aqueous solution 
are hydrated. It is necessary to know the extent of this hy- 
dration to fully understand electrolytic solutions* Unfortunate- 
ly the common use of the hydration concept implies a greater 
understanding than we now possess. Too often hydration is used 
as a convenient excuse for anomalies in theories of solutions. 

II. Ionic Hydration Theories 

Early hydration theroies pictured the ions as being chem- 
ically bonded to the water molecules of hydration. Then f rom 
1920 to 191+0 the success of the Debye-Huckel theory in account- 
ing for properties of dilute electrolyte solutions led to the 
adoption of a more physical than chemical interpretation of ion- 
solvent interactions. The energy of this interaction was given 
in terms of the bulk dielectric constant of the solvent. However, 
it soon became apparent that the use of the macroscopic dielectric 
constant could not suffice for all energy calculations in more 
concentrated solutions, and a molecular approach was necessary. 
Recognition of this need plus new methods of investigation have 
revived interest in the hydration of ions. 

Most current theories refer to "primary" and "secondary" 
hydration to differentiate the water molecules directly adjacent 
to the ion from the bulk of the solvent. 1 Robinson and Stokes 2 
show that only a small polyionic ion is likely to exert appreci- 
able orienting effects on a second sheath of water molecules. 

Some theories attempt to account for all of the hydration 
interaction and its effects by considering only the primary hy- 
dration. In such a theory Buckingham 3 assumes an immobilized 
layer of four or eight water molecules surrounding the ion, and 
calculates the interaction energies of these molecules. The 
remainder of the solvent is considered as a continuous medium. 

One of the more popular models, advocated by H. S. Frank, 4 ' 5 
gives more emphasis to the secondary hydration. Frank proposes 
a model in which the ion is surrounded by a sheath of compari tively 
immobilized water molecules. Between this sheath and normally 
structured water exists a region of "structure-breaking" water. 
This structure-breaking region is due to a balance between two 
orienting tendencies: the normal structure in water and the 
orientation of the dipole in the strong field of the ion. The 
extent and total effect of hydration is considered a reflection 
of the relative sizes of these regions. 

Samoilov 6 stresses that the assumption of definite stoichio- 
metry is only a crude approximation, and supports a model of 



continually exchanging water molecules. Energy calculations are 
based on the action of the ions on the trans lati onal motion 
of the nearest water molecules, 

III.. Proposed Methods of Measuring Hydration Numbers 

Various methods are based on transport pheonomena: 

A, The difference in the amounts of water transported by 
the anions and cations can b e calculated from a modification of 
the Hittorf method for mearsuring transport numbers, 2 

B* Diffusion coefficients of concentrated electrolytic 
solutions are interpreted in terms of the sum of the hydration 
numbers of the two ions, 2 

C* Ionic mobi li ti es are also interpreted in terms of hy- 
dration numbers. The advantage of this method is that values 
are obtained for individual ions. 1 

Numerous methods use thermodynamic calculations to estimate 
the effect of hydration on equilibrium properties: 


D, The solubility decrease of a non-electrolyte (salting- 
„ut effect) caused by the addition of an electrolyte reflects 
the amount of free water which is "tied up" by the added ions* 7 

E* Deviations of the solutions from Raoult's law and the 
Debye-Huckel theory are used to calculate hydration numbers, 2 
However, this method attributes all deviation to hydration, 
whereas other inestimable factors undoubtedly a 1 so contribute, 

F* The rapid rise of the acidity function of strong acids 
with increasing concentration is attributed to hydration of the 
proton beyond ^O" 1 "* 6 

G, Entropies of dissolved salts are obtained rather easily 
from experimental data, and are used to determine the hydration 
trends for families of ions* 4 

Two direct approaches should be mentioned: 7 

H* The effect of the addition of electrolyte on the total 
volume of solution is used as a measure of hydration, 

1, The assumption that the water molecules immediately 
surrounding the ions are at maximum density serves as a basis 
in using compressibility data for estimating hydration numbers. 

Isotope exchange, 7 Raman spectra, 9 nuclear magnetic resonance, 10 
x-ray spectra, 11 ' 12 ultra violet spectra, 13 surface properties, 14 
and rate measurements 15 are also used to give information on 
the hydration of ions, 

IV* Conclusion 

The term "ionic hydration number" implies a simple stoichio- 
metric interaction, an idea which is valid for only a few ions* 
As a result, the energy of hydration is a much more meaningful 
quantity than is the hydration number. At present several physical 
pictures of the interaction can b e used to explain the effects 
of hydration* 


■ ■ 


rfT . 


• J 






V* References 

U J* 0„ Bockris, Quart*. Revs*, 3, 173 ( 1914-9) - 

2. R # A. Robinson and R. H. Stokes, "Electrolyte Solutions", 
Academic Press, New York, 1955* 

3» A, D. Buckingham, Disc. Faraday Soc, 2Lj., 151 (1957). 

1|. H. S» Frank and M. W. Evans, J. Chem. Phys., JJ3, 507 ( 19*4-5) • 

5. H. S. Frank and V/ en-Yang Wen, ^isc Faraday Soc, 2l±, 133 

6. # Y. Samoilov, ibid, 2ij., ll|l (1957). 

7. R. P. Bell, Endeavor, J/7, 31 (1958)* 

8.. D. M. Wyatt, K. N, Bascomb, and R. P. Bell, Disc Faraday Soc, 
glfc, 158 (1957). 

9. T. F* Young, ibid, 2lfc. 37 (1957). 

10» O. Redlich, ibid, 2l±, 87 (1957). 

11* C. L* Panthaleon van Eck, H. Mendel, and W. Boog, ibid, 
2Lt, 201 (1957). 

12* J. D. Bernal and R. H. Fowler, J. Chem. Phys., 1, 515 (1933). 

13. M» Smith and M. C. R# Symons, Disc. Faraday Soc, 2Lj., 206 

lLj.. J. E. B. Randies, ibid, 2^, 19l| (1957). 
15. M. Eigen, ibid, 22j., 25 (1957). 

« , 





Catalysis of Organic Reactions by Metal Carbonyls 

John G. Verkade December 9, 1958 

The metal carbonyls of the first transition series are the most 
important and interesting members of this group of compounds from 
the point of view of both applications in organic chemistry and the 
existence of complexes containing metal, carbon monoxide and organic 
groups in the same molecule . 

The reaction between metal carbonyls and acetylene was dis- 
covered by Reppe (l) in 1939. Catalytic quantities of nickel car- 
bonyl were used to effect the synthesis of acrylic acid and its 
derivatives from acetylene, carbon monoxide and water. Amines and 
alcohols can also be used instead of water as active 

HC = CH + CO + H 2 


CH 2 = CHCO2H 

hydrogen compounds to produce the amides and esters, respectively. 

In the discussion of the mechanisms of two of the type reactions 
listed below, special attention will be devoted to the nature of the 
complexes, which presumably form as intermediates. 

Type Reactions Catalyzed by Metal Carbonyls 

1. Hydroformylation (l) ( "oxo " reaction). A hydrogen and a 
formyl group are added across a double bond in the presence of 
cobalt octacarbonyl. 

Co 2 (CO) 8 f^^^tvL 

+ (CO + H 2 ) * 



2. Hydrohydroxymethylation (2). In this reaction a hydrogen 
and a hydroxymethyl group are added across a double bond. This re- 
action is believed to occur in two steps: (a) a hydroformylation 
and (b) a reduction of the aldehyde group to an alcohol. The mecha- 
nism of this reaction differs in several ways from the hydroformy- 
lation reaction (a) in that Fe(CO) 5 is used in place of Co 2 (CO) 8 as 
the catalyst, (b) the reaction is carried out in aqueous base rather 
than in organic solvents and (c) water instead of hydrogen gas 
serves as the source of H atoms. 

3. Hy decarboxylation (l). This reaction is carried out using 
Ni(CO) 4 as shown in the Reppe reaction (vide supra). 

4. Hydroesterification (3). This reaction is also carried out 
under Reppe reaction conditions except that Co 2 (CO) 8 and alcohol are 
used instead of Ni(CO) 4 and water. 

5. Hydrooyanation (4). Hydrogen cyanide is added across a 
double bond in the presence of Co 2 (CO) 8 . Little is known of the 
intermediate products . 


■■'.■ ■ . • \ . • 

■.. . ■ • '. 

• i 



-'■■..■' . . . 



. ■ .... 




The Hydroformylation Reaction 

1. Mechanism and Intermediates in Olefin Reactions. 
On the basis of the difference in rates of reaction with varying 
carbon monoxide and hydrogen pressures, it was postulated (5,6,7) 
that the first step in the reaction entails formation of the com- 
plex (I) between the catalyst and the olefin. On the basis of the 
kinetics, Wender (l) postulated that I reacts 

+ CO + H; 

Co 2 (CO) 

[Co 2 (C0) 7 .C 6 Hi o ] + CO -> 

with H 2 to., form a second complex II (i.e., H 2 [Co 2 (CO) T C 6 Hio] ) . Some 
possible structures of I and II will be discussed in the light of 
experimental evidence (8,9). It is found that H Co(C0)4 plays an 
important role in these mechanisms (9>10,ll). Support for the ex- 
istence of II arises from olefin isoraerization experiments (l). 

2. Mechanisms and Intermediates in Acetylene Reactions. 
Tirpak et al (12) followed the reaction kinetics of hexyne-1 and 
hexyne-2 with Co 2 (CO) 8 by measuring the rate of CO evolution. Their 
interpretation of the mechanism is based on the possible existence of 
a stable and an unstable form of Co 2 (CO) 8 in equilibrium. Possible 
structures are discussed for the postulated intermediates in the 
simplest mechanisms fitting all the experimental data. 

The Hydrohydroxymethylation Reaction 

The importance of the [H 2 Fe 2 (CO) 8 ] = ion and the H 4 Fe 2 (CO) 8 
molecule is emphasized in their roles as isomerization and reduction 
catalysts. The fact that the hypothetical intermediates in this re- 
action are isoelectronic with those in the hydroformylation reaction 
lends support to the similarities in the reaction mechanisms and 
products . 

Recent Work on Organo -Metal Carbonyl Complexes 

That the postulation of structures for intermediates in these 
reactions is dangerous and misleading is clearly demonstrated by the 
X-ray determination of the structure of [H 2 Fe 2 (CO) 8 -CHsCsCCHs] (13). 
The structure is quite different from any of the proposals made 
heretofore (14,15,16). 

Jonassen (17) has found that a complex between HCo(CO) 4 and 
butadiene forms which is diamagnetic, has no acidic hydrogen and is 
monomeric . A possible structure is not given although an X-ray in- 
vestigation is now in progress. 

Sternberg (18) has postulated a structure for the complex 
formed between Fe(CO) 5 and CH3C-CCH3. The suggested structure con- 
tains a quinone. If correct, this would be a molecule in which the 
carbon skeleton is formed during the reaction. Duroquinone is 









The use of structures of intermediates postulated on the basis 
of evidence excluding X-ray determination must be limited to under- 
standing the course of the reaction only. It will be recognized 
that the actual structure may be very different from that postulated, 
It must also be realized that the interpretation of the mechanisms 
of these reactions may undergo radical alterations as further 
structure determinations of such intermediates are carried out. 


1. I. Wender and H. Sternberg, Project 1567* Central Experimental 
Station, Div. of Solid Fuels Technology, Bruceton, Pa., Sept. 
7, 1956. 

2. W. Reppe and H. Vetter, Ann., 582 , 133 (1953). 

3. R. Ercoli et al., Chim. e Ind., 37, 865 (1955). 

4. P. Arthur et al., J. Am. Chem. Soc, j6, 5364 (1954). 

5. G. Natta and R. Ercoli, Chim. e Ind., 34, 503 (1952). 

6. G. Natta and R. Ercoli, J. Am. Chem. Soc, 76, 4049 (1954). 

7. H. Greenfield et al., Abstracts of Papers, 126th Meeting of 
the ACS, New York, N.Y., Sept., 1954. 

8. H. Sternberg et al., J. Am. Chem. Soc, £6, 1457 (1954). 

9. H. Adkins and G. Krsek, J. Am. Chem. Soc, 70, 383 (1948). 

10. I. Wender et al . , J. Am. Chem. Soc, 71, 570T (1951). 

11. I. Wender and H. Greenfield, ibid., £57 4079 (1952). 

12. M. Tirpak et al,, ibid., 80, 4265 (1958). 

13. A. Hock and 0. Mills, Proc Chem. Soc, 233 (1958). 

14. H. Sternberg et al., J. Am. Chem. Soc, 78, 3621 (1956) . 

15. F. Clarkson et al., ibid., £8, 6206 (195^7. 

16. H. C. Longuet -Higgins and L. E. Orgel, J. Chem. Soc, 1969 

17. H. Jonassen et al., J. Am. Chem. Soc, 80, 2586 (1958). 

18. H. Sternberg et al., ibid., 80, 1009 (1958). 



Mitsuru Kubota December 16, 1958 

I. Introduction 

Infrared spectro copy has been firmly established as a method 
for the elucidation of molecular structure chiefly by the empirical 
correlation of absorption bands to certain atomic groupings and 
more recently by the study of absolute intensities of these bands* 
Since the infrared spectrum reflects vibrational energy transitions 
within the molecule (intramolecular) as well as between molecules 
( intermol ecular), it is not at all surprising that spectral 
variations occur upon changing the state of a sample* In the 
transition from gas to liquid to solid, variations in spectra 
such as differences in relative intensities and positions of bands, 
splitting of broad bands into several sharp components and apparent- 
ly new bands appearing in one state, absent in the others, are 
often detected* 1 ' 2 

II* Experimental methods 

The primary difficulty in obtaining solid spectra is due to 
the loss of incident radiation by reflection and scattering* 
Quantitative studies often encounter difficulties in controlling 
sample thickness, and in evaluating particle size and particle 
size distribution. 1 Among the more common methods which are 
employed are the following: 

A* Mul 1 inq techniques . The solid sample is ground in an 
agate mortar, and is mixed with high boiling petroleum (Nujol)* 
The mull obtained is then placed between salt plates* Adequate 
coverage of the spectrum is attained by the use of other mulling 
materials such as fluorinated hydrocarbon (fluorolube) or hexachloro- 
butadiene* 1 

B» A 1 ka 1 i metal halide disks , A small quantity of sample is 
ground and mixed with 150 to 200 times its weight of finely 
powdered KBr or other alkali metal halide* The mixture is placed 
in a special die, evacuated and pressed for several minutes to 
yield a clear pellet* 4 " 6 

C» Single crystals . The spectrum of a crystal can be taken 
directly, or if necessary it can be cleaved prior to analysis* 7 
Small crystals can also be studied by mounting them at the focus 
of a reflecting microscope*© 

D* Sol id f i 1ms , The spectrum of a solid can be obtained by 
pouring its melt on a supporting plate upon which it solidifies* 
More often a solid film is obtained by subliming or crystallizing 
the solid upon a cold plate* 9 Better resolution is often obtained 
by "tempering" the solid film* 10 In the matrix isolation method 
the gaseous sample and suitable inert matrix is sprayed on a cold 

II* Some characteristics of infrared spectra of solids 

Except in instances where hydrogen bonding is involved, the 
spectra of solids obtained with low resolving instruments show only 






. '- 

. ■ . 


■ [ 





■ - . . 









minor variations* 11 Greater resolution however, shows a wealth 
of fine components, which are markedly different from those of the 
liquid or gaseous states. Much current research is focused at 
the lack of satisfactory explanations for many spectral features 
of the solid state. Some aspects of solid state spectra which 
have been fairly well demonstrated are to be discussed,. 

As a consequence of the close order crystal env i r*nmen.t , inter- 
molecular effects are greatly enhanced in the solid state. With 
polar compounds dipole-dipole interactions often lead to spectral 
changes. Induced absorptions which do not highly depend on 
polarity may persist, in which case they are intensified, or may 
vanish when a liquid is crystallized* Such absorptions have been 
ascribed to dipole moments induced by. intermoleculer forces in 
clusters of interacting molecules* 1 E * -* 

Splitting of bands throughout the crystal spectrum result from 
combination bands between internal modes and lattice torsions. 
This effect, also known as librational interaction, can account for 
most of the complex fine structure of an absorption band. 14 ' 15 First 
order theory and methods for formulating selection rules in terms 
of crystal symmetries have been developed and well demonstrated* 1 6 * : 7 
Recent low temperature studies highly support this theory of 
librational interaction* 6 ' 16 

Alterations in crystal lattice, such as changing from one 
polymorphic form to another, which lead to changes in the immediate 
environment of vibrating groups, often cause spectral changes* 
Rotational isomers which exist in the liquid and gaseous states 
are often eliminated in the solid state* 1 ' 11 

Orientation effects become more pronounced in the solid state 
due to the increased orderly arrangement of molecules* If the 
molecules are lined up in a manner such that the electric vector 
associated with some localized group vibration lies along the axis 
of-a partially polarized beam of radiation, the expected absorption 
band can be weak or even absent* 1 

The Christiansen filter effect (anomolous reflection and 
transmission) often occurs in molecular and Ionic crystals. Peaks 
associated with this effect are much closer together for molecular 
crystals, since they are produced by a single frequency instead 
of a wide band of lattice frequencies* 19 

IV* Some applications* 

A* Chemi sorbed molecules . Infrared investigations of molecules 
chemisorbed on solids give information on the structures of 
intermediates involved in heterogeneous catalysis* Radiation 
losses by scattering and absorption are minimized by dispersing 
the small metal particles in supporting silica or alumina, or by 
using evaporated metal films. Carbon monoxide adsorbed on 
platinum films exhibits a band at 2050 cm" 1 similar to that in 
nickel carbonyl, thereby suggesting a "linear type" Pt-G-0 bond* 
Carbon monoxide adsorbed on silica supported palladium shows 
peaks at 1820 and 1925 cm, similar to peaks in Fe 2 (CO)g. This 
suggests "bridge type" linkages. 20 Eischens concludes that such 
infrared studies provide a precise means of distinguishing 













between chemi sorption and physical adsorption* 21 

B* Matrix Isolation . Leisurely spectral study of extremely 
unstable species, even free radicals, can be accomplished by this 
method. 22 That mechanisms of reactions can often be elucidated, 
is shov/n vividly in a study of the photolysis of nitromethane and 
subsequent detection of nitroxyl. 23 Spectral studies of hydrogen 
bonding effects have been illuminated by the matrix isolation 
method as a consequence of reduced band widths which aid in the 
detection of overlapping and accidentally degenerate bands. 24 

C. Structural studies wi th polarized radiation . Quantitative 
treatment of results of polarized infrared studies can often com- 
plement structural information obtained by x-ray diffraction* 
Using this method, along with isotopic substitutions, L.H. Jones 
has worked out complete fundamental assignments for solid complex 
cyanides, such as KAu(CN) £ , KAg(CN) 2 and K 3 Cu(CN) 4 » Calculations 
of stretching force constants indicate that the metal is bonded to 
the carbon rather than the nitrogen atom. 25 


. . . 

j ■ ._ . 

" . - ' 

not • .H« ' 

j j . . ■ ' 

: , ' - :■;■ .... 

' '. I ■ . • . 




R,. N. Jones and C. Sandorfy, "Chemical Applications of Spectro- 
scopy", pp. 29l|-32o, Interscience, New York (1956). 
R. E. Richards and H. W. Thompson, Proc f . Roy. Soc. (London), 
A195, 1 (19U8). 

J. S. Ard, Anal. Chem. 25, 17I4.3 (1953)- 
A. W. Walker, J. Phys. Chem. 61, I4.50 (1957). 

M, M, Stimson and M. J. 0*Donnell. J. Am. Chem, Soc. 7k* l8°5 
(1952). • "* 

T. J, Lane, D. N. Sen and J. V.. Quagliano, J. Chem. Phys. 22, 

1855 (1951*3. "~ 

W. R. Busing and H. W. Morgan, Ibid, 28, 998 (1958). 

L. H. Jones, Ibid. 21, 1891 (1953). 

A, Terenin. Vi. Filimonow and D. Bystrow, Z. Electrochem. 62, 

180 (1958). — 

R. C. Millikan and K. S. Pitzer, J. Am. Chem. Soc.. 80, 3515 

(1958). ~"~ 

L. J. Bellamy, "The Infra-red Spectra of Complex Molecules", 

PP. 377-38Ij-, Methuen and Co. Ltd., London (1958). 

R. S. Halford and I. Ichimasa, Ann. Rev. Phys. Chem. 7, i\2$ 


M. K. Wilson and V. A. Crawford, Ibid. 9. 3kk (1958). 

D. A. Dows, J. Chem. Phys. 29, ifik (19^H3. 

D. F. Hornig, Disc. Faraday Soc. 9, 115 (1950)» 

R. S. Halford, J. Chem. Phys. ljL 8 ( 19U-6 ) • 

D„ F. Hornig, Ibid. _16, IO63 (l9lj-8)» 

R. M. Hexter, Spectrochim. Acta, _10, 28 1 (1958). 

W. C. Price and K. S. Tetlow, J, Chem. Phys. 16, 1157 (l9lj-8). 

R. P. Eischens, J. Chem. Education, J5, 385 (1958). 

R. P. Eischens and W. A. Pliskins, Advances in Catalysis, 10, 

1, (1958). 

G. C. Pimental, Spectrochim, Acta, _12, 91+ (1958). 

H. W. Brov/n and G. C. Pimental, J. Chem. Phys. 29, 883 (1958). 

M. Van'Thiel, E. D. Becker and G. C. Pimental, Ibid. 22, 95 


L* H. Jones, Ibid. 27, 268, 665, 1578 (1957); 29, U-63 (1958). 

■....■_ .. t bn 

... - r - ■ > * 

, . . . „ i « ■ .. '. : •'.... . . T 


' ■ . • ■ . . . 

, . • «C ...■•■». 


• ■..-., ■■ ■ . : 

t, ' ' - 


r * 

l) , ,-g ,t « ... 

» L | * » ■ • . 

. ' ■ , , , . 




C. H. Travaglini January 1J>, 1959 

I. Introduction 

The theoretical approach to surface catalysis was first con- 
sidered in a series of chemical papers in the early 1930's by 
Langmuir, who suggested that adsorption occurred by chemical forces . 
Solid state theories can be used as a first approximation to corre- 
late catalytic activity with solid state behavior, if the differences 
of bulk solid and surface are considered. The nature of the adsorbed 
film on metals, the reaction path and activated complexes involved 
will be examined, and related oxide studies considered. 

II . The Solid and the Activated Complex 

Catalytic influence may be divided into two factors, geometric 
and electronic, (l) The geometric factor involves fitting the 
substrate (adsorbate) to either one atom (single -point adsorption) 
or several points simultaneously (Multiplet Theory) of the catalyst 
surface. (2,3) The two factors cannot strictly be separated, since 
the bond length between the two surface atoms is characteristic of 
the bond concerned, but attention here will focus on the electronic 

The two main theories of metals help correlate the electronic 
factor with catalytic activity, which, in a metal, depends on the 
electronic work function, electron density gradient, and the electron 
density at the Fermi surface. (4,5) The energy band theory of Bloch 
and Brillouin visualizes the valence electrons of metals moving free- 
ly through a lattice of positive cores, providing a periodic po- 
tential field, arising at the surface. The energy of the electron 
remains quantized, but overlapping energy levels form allowed energy 
bands, separated by gaps of non -allowed energy, limited by the energy 
of the Fermi surface (highest filled electron level at 0°K.) (6) 
Electron conductivity occurs if a band (Brillouin zone) is partly 
full, as only an electron in a singly occupied level can migrate in 
an external field. If the band is full or the energy gap between 
zones is too great, semiconductors or insulators result. In the 
transition metals, the nd and (n+l)s levels broaden and overlap, 
leaving unfilled holes in the d-band of approximately .6/g-at. 

The Pauling resonating -valence -bond theory describes a metal in 
terms of hybrid electron -pair bonds, resembling ordinary covalent 
bonds. (7*8) Metallic properties are based on possession of all or 
some atoms in a given metal of a free orbital which allows uninhibited 
or unsynchronized resonance, as In Li -Li Li^Li . Some d- orbitals 

Li -Li ' Li^L'i 
in the transition metals remain unhybridized, as atomic orbitals, 
which can be used in cheraisorption. Here again ferromagnetic 
saturation moments reveal .6 unpaired d electrons per gram-atom. 

The energy of bond formation may be calculated using Lennard- 
Jones potential energy curves. (9* 10) For a maximum free energy 
change, the substrate particles must be In a geometric site with 
miniraun activation energy necessary for adsorption, and with the 
binding electrons in suitable electronic orbitals, as in the crystal- 
lite planes of least packing density, 111 or 110. (ll) The rates of 






. J 



. ■ ■ 



i G 


heterogeneous chemlsorption reactions, controlled by the formation 
of a bound complex, are examined using the Polanyi profile method: 

Reactants -> Adsorbed ■*• Adsorbed ■+ Products , 

Reactants Products 

assuming 2-*\5 as the rate determining step. Examination of these 
curves reveals factors for decreasing the energy of activation and 
hence increasing rate. 

III. Chemisorption and Chemical Reactions on Metals and Oxides 

Early experimental work by Schwab on formic acid decomposition 
on alloy catalysts indicated a link between the energy of activation 
and the energy necessary to remove an electron from the substrate to 
the Fermi level of the metal. (12) By adding mult-valent metals to 
monovalent ones, thereby increasing the energy of the Fermi surface, 
the activation energy for decomposition increases as expected. (4) 
Hydrogenation catalyzed on transition metals revealed this same de- 
pendence on atomic d- orbitals. Eley and Couper found in their in- 
vestigation of the para-H conversion on Pd-Au alloys that the empty 
d-orbitals could come to the catalyst surface, accept electrons, and 
form low energy complexes. (13) Oxide semiconductors of the n-type 
favor H 2 , CO and substances which donate electrons, while p-types 
favor electron-acceptor substrates, such as N2O. (1*0 In general, 
electron excess lattices favor dehydrogenation, but electron defect 
lattices favor dehydration. (10) 

Chemisorption involves metal half -bonds, similar to those in 
electron-deficient compounds, so that depending on the crystal 
lattice and plane exposed, single or multiple bonds (M-H, M=0, M=N) 
may form. (19) Trapnell found that all sp and dsp -bonding metals 
adsorb 2 except gold, and only dsp metals adsorb N 2 , H 2 , CO, C 2 H 4 
and C 2 H 2 . (8) The heat of chemisorption as calculated by Eley de- 
creases as the fraction of surface covered increases. (6, 15) The 
field emission microscopy studies of Gomer reveal a primary mobile 
film and a secondary stationary film of adsorbate, with different 
energies of activation. (16) 

Suhrmann's examination of electronic work functions as a 
measurement of electronic interaction on a metal surface shows that 
2 , N 2 , N 2 and CO require surfaces with low work functions (easily 
release electrons), but H 2 adsorption favors a high work function. (21) 
Even saturated hydrocarbons adsorb at low temperatures on metals by 
electron release as shown by the decrease in work function. 

For the elucidation of mechanisms for catalytic reactions on 
metals, the energy of activation and frequency factor (E, A of 
Arrhenius) or heat and entropy of activation (of Eyring) must be 

considered. (26) This mechanism j§ + p-H 2 ■* H'' "*H ■* o-H 2 + S 









EO X o 










... .- " 


for the para-H conversion on W and Pd-Au is consistent with the A E+ 
involved. (13, 17, 18) Abruptly at .6 atomic percent of Au in Pd, 
the AE-t increased, perhaps indicating a change to the Bonhoeffer- 
Farkas mechanism, 2M-H ■* M-H + o-H 2 , simultaneously with the filling 
of the d- orbitals. (20) Dowden and Reynolds support the relation- 
ship of activity with the number of d-holes with their dehydrogenation 
work on binary solid solutions of Group VIII and lb elements. (4) 
The efficacy of styrene dehydrogenation on a Ni-Cu alloy series de- 
creases as the number of d-holes approaches zero, as the Cu concen- 
tration increases. Further correlation comes with Trapnell's results 
indicating the transition metals W, Ta, Mo, Ti, Zr and Fe as more 
active for cheraisorption. (22) 

Adsorption on oxide surfaces involves electron transfer between 
the adsorbate and defects near the surface. Water often helps promote 
reactions, especially on insulators, such as A1 2 3 . Propylene 
isomerization reaches a maximum at 85$ Si0 2 with A1 2 3 , the point of 
maximum acidity (greatest H-ion activity). (23) Heating in hydrogen 
or adding Ga 2 3 increases ZnO activity, possibly by formation of 
surface anionic defects, such as OH". Oxygen lattice exchange 
studies on Cu and Zn oxides by Winter show a lack of equilibration 
and oxide formation, suggesting that reaction occurred on interstitial 
metal atoms, which then may have diffused into the catalyst interior 
rendering the dissociation irreversible. (24) The oxidation of S0 2 
by metal ions also depends on the ability of the cation to transfer 
electrons, possibly via a radical -ion mechanism. (25) 

IV. Conclusion 

For the type and intensity of electron interactions between 
metal surfaces and adsorbed molecules, the factors which must be 
considered are the magnitude of the work function, the electron af- 
finity of the species to be adsorbed, the asymmetry of the electronic 
configuration, the presence of unpaired and vr-electrons, and the 
effects due to steric hindrance. (21) Changes in the activation 
energy for adsorption depend on the electron concentration, or on the 
degree of completion of the Brillouin zone, indicating that the main 
process in heterogeneous catalysis is the formation of adsorbed 
layers by electron sharing. 




1. D. D. Eley, Z. Elektrochem. 60, 797 (1956). 

2. R. H. Griffith, 'Advances in Catalysis, " Vol . I, 91 (1948). 
3- R. E. Kirk and D. F. Othner (Ed.), Encyclopedia of Chemical 

Technology, Interscience Encyclopedias, Inc., New York, 
New York, Vol. 35, 1949, p. 245. 

4. D. A. Dowden and P. W. Reynolds, Discuss. Faraday Soc. No. 8, 
184 (1950) . 

5. W. Hume -Rothery , "Atomic Theory for Students of Metallurgy," 
Institute of Metals, London, 1946. 

6. D. D. Eley, Reilly Lectures, Vol. VII (195*0, Notre Dame, Ind. 

7. L. Pauling, Proc . Roy. Soc. [London] Ser. A I96 , 343 (19^9) - 

8. Baker and Jenkins, ''Advances in Catalysis," Vol. 7, 1 ?1955) • 

9. J. E. Lennard -Jones, Trans. Faraday Soc. 28, 333 (1932). 

10. W. E. Garner, Discuss. Faraday Soc. No. 8, 246 (1950) . 

11. D. A. Dowden, J. Chem. Soc. [London] 1950 , 242. 

12. G.-M. Schwab, Trans. Faraday Soc. 42, 689" (I9 ] l6). 

13. D. D. Eley and A. Couper, Discuss. Faraday Soc. No. 8, 172 (1950) 

14. C. Wagner, J. Chem. Physics 18, 69 (1950). 

15. D. D. Eley, Discuss. Faraday Soc. No. 8, 34 (1950). 

16. R. Gomer, 'Advances in C atalysis, " Vol"." J_ t 93 (1955). 

17. D. D. Eley and A. Couper, Proc. Roy. Soc." [London] Ser. A 211 , 
544 (1952). 

18. D. D. Eley and E. K. Rideal, Proc. Roy. Soc. [London! Ser. A 178, 
429 (19^1). 

19. R. E. Rundle, J. Amer. Chem. Soc. 6_2, 1327 (1947) • 

20. D. D. Eley, Quart. Rev. 3, 209 (19^9). 

21. R. Suhrmann, "Advances in Catalysis, " Vol . 7, 303 (1955). 

22. B. M. W. Trapnell, Proc. Roy. Soc. [London] Ser.' A 218 , 
566 (1953). 

23. A. Clark, Ind. Eng. Chem. 45, 1476 (1953). 

24. E. R. S. Winter, J. Chem. Soc. [London] 1955, 3824. 

25. M. Boudart, Ind. Chim. Belze 23, 383 (1958). 

26. E. Creraer, "Advances in Catalysis, ,: Vol . 7, 75 (1955)- 



L. C. Thompson February 10, 1959 

I. Introduction 

The thermal decomposition of the nitrogen oxides is of great 
interest for several reasons. The primary reason may be summed 
up in the following quotation (l): "Although examples are fewer, 
the oxides of nitrogen with oxygen atoms, molecules, and ozone 
illustrate almost as many principles of reaction kinetics as do 
all gas-phase organic reactions. *hese reactions form a family 
of their own with many elementary reactions appearing in more than 
one system." 

Another extremely important reason is the fact that the phenom- 
enon of nitrogen fixation has not been completely explained. Work 
on the kinetics of the thermal decomposition of the nitrogen oxides, 
coupled wi th photochemical studies, promises to give an insight 
into this problem. 

II.. Bond "Energies" 

In any kinetics study a knowledge of the bond energies of 
the reactants and products is helpful. Bond energies for some of 
the nitrogen oxides are listed in Table I. 

Table I 

Bond Energies 

Compound B(N-O) kca ls/mole B(N-OM) kca ls/mole B(N-N) kca ls/mole 

NO 123 

N0 2 98 

N 2 4 99 13.8 

N0 3 82 

N 2 5 90 21.6 

They are defined as follows: 

a x( g) (g) K Q) 

B(N-O) = oH a = 1 

- T a°x 

with appropriate modifications for a ^ 1* 

Using these bond energies, one can predict qualitatively the 
path of reaction when two or more alternative paths are given. 

'. . 


The decomposition of N 2 5 into N0 2 and N0 3 gives an example of the 
use of bond energies. If we assume that the structure of N 2 5 
can be written as 


-.. N-O-N 

we can picture the decomposition as the breaking of the N-O-N 
bond in an appropriately excited state. The activation energy 
involved in this reaction should be roughly equival ent to the 
energy involved in the breaking of the bond. Such is, indeed, 
the case (2). 

In other cases postulated mechanisms may be ruled out if 
the energy of activation is knovji and one can assume some reasonable 
value for the energy of the bond to be broken. If these two are 
widely different, the proposed mechanism can usually be discarded. 

III. General Experimental Methods 

A few of the many experimental techniques v/hich have been 
used in this field are discussed below with particular reference 
I to the methods for determining concentrations. 

|A„ Direct Pressure Measurements* 

The classical study of the decomposition of nitrogen pentoxide 
[is that of Daniels and Johnston and of Daniels and other co-workers(3 ) • 
This reaction obeys the stoi chiometry 

2N 2 5 -> 2N 2 4 + 2 

The important feature of the apparatus employed in this study was 
the device used to measure the pressure within the reaction vessel. 

.A platinized glass diaphragm separated the reaction vessel from 
an outer space, the pressure in v/hich could be measured by a mercury 

l manometer, and could be adjusted to a predetermined pressure at 
will. When the pressure in the reaction vessel was equal to, or 
greater than that outside, the diaphragm made electrical contact 
with a platinized glass knob. he pressure on the outside of 
the diaphragm was increased, and the time for the inside pressure to 
reach the new value and cause a deflection of a galvonometer was 
noted with a stopwatch. This procedure was followed until no 
further change was noted, ^he dissociation of N 2 4 somewhat 
complicates the pressure readings, but it can be taken into account. 

B. Spectroscopic Methods: 

1. Kaufman and co-vo rkers have used the following technique 
I in the decomposition of nitric oxide (I4.). In runs where the % 
decomposition v/as greater than 30%, excess oxygen was added to 
convert all of the remaining NO to an equilibrium mixture of n 
!D2 -N 2 4 . When the % decomposition was small, excess NO was added - 
to convert all the oxygen formed in the decomposition to an equili- 
brium mixture of N0 2 , N 2 4 and N 2 3 . T he samples containing 






excess oxygen were analyzed at 3CCO and 3200 R since the absorption 
coefficient of N 2 4 is "roughly" tv/ice that of N0 2 and no account 
need be taken of the N0 2 -W 2 4 equilibrium. Samples containing 
excess NO were analyzed at 31+50, 3500, 3550, and 36OO R so that the 
variable N 2 3 could be taken into account, 

2. Hisatsune, Crawford, and Ogg have followed the NO-N 2 5 
reaction using a fast-scanning spectrometer (5)» ^his instrument 
can be used if the reaction has a half-time of about one second 
when two neighboring bands of different molecular species are 


C. Shock Tube Method 

The shock tube pyrolysis of N 2 5 has been reported by Schot t 
and Davidson (2). *n this technique a shock wave is produced in 
N 2 5 by argon and the N0 2 and N0 3 produced are measured spectro- 
photometrically. *he time required for the dissociation of N 2 5 
is of the order of 5 '1 sec at 600° K. ^he subsequent decomposition 
of N0 3 , uhich is slow compared with the dissociation, can also be 

IV. Examples: 

A. Thermal Decomposition of N0 2 : 

The thermal decomposition of N0 2 was first investigated by 
Bodenstein in 1922 (6). He found that the reaction followed 
second order kinetics and could be written as: 

N0 2 + N0 2 -♦> NO + NO + 2 k 

-d(NO? ) = k (NC 2 ) 2 
dt l 

Measurements in 1956 by Rosser and V/ise (7) using a spectro- 
photometry method of analysis and concentrations in the range 
10-7 mole/cc. ( - 6 mm Hg) confirmed the earlier work and gave 
the following expression for k.: 

k = 10 12 * 6 exp(-26,900/RT cc mole"* 1 sec" 1 

However, Ashmore and Levitt, working at lower pressures (•'-._3«3 nun Hg) 
near I^OOoc, have found anomalously high initial rates for the 
decomposition (8). ^he slope of a plot of l/(N0 2 ) vs. time falls 
steadily and then remains constant for a considerable time. The 
same result was obtained when higher initial pressures were used, 
but the limiting slope was attained more rapidly. This non- 
linearity was completely suppressed by the addition of an equal 
pressure of NOy but N 2 up to 0,5 atmosphere had no effect. Ashmore 
and Levitt proposed the following mechanism: 


N0 2 + N0 2 -> NO + NO + 2 k x 

N0 2 + N0 2 -> NO + N0 3 k 2 

N0 3 + NO ~> NO 2 + N0 2 k 3 

N0 3 + N0 2 -> NO + N0 2 + 2 k 4 

Ml of these steps except (2) have been invoked in a discussion of 
the kinetics of the N 2 5 decomposition or reactions of the pent- 
oxide. Assuming the steady state treatment for N0 3 , the following 
is obtained: 

d l/(NOg) = -4 d(NO ? ) sr (k + k 2 ) -k ? k 3 (NO) (5) 

dt ( N0 2 P dt * kJTNO) + k 4 (N0 2 j 

Jsing this rate law it iS',§^en that-.early in: jbhe. i?eaction when^ (NO) 
Is small, plots of l/(N0 2 ) vs. time will have slope (ki + k 2 ) j and 
that later, if k 3 x> k 4 and k- ,:• k 2 , k 3 (NO) rapidly becomes much 
greater than k 4 (N0 2 ) and the slope becomes k*, which can be identi- 
fied with the normal value for the rate constant of the decomposi- 
tion. T he effect of added NO is to make k 3 (NO) -?> k 4 (N0 2 ) from the 
beginning. The mechanism is also consistent with the observation 
that the addition of N 2 has no effect on the initial rate. 

3. T nerraa 2 decomposition of N 2 5 (9)» 

The study of the decomposition of nitrogen pentoxide has 
Dlayed an important role in the development of gas kinetics. 
*irst investigated by Daniels and Johnston(3), this decomposition 
vas, until 1925, the only accepted example of a first-order gas 
reaction. The reaction was subsequently studied over a wide var- 
iety of temperatures and pressure by other workers using many 
iifferent techniques. Th ese investigations were in substantial 
igreement with Daniels and Johnston T s results. 

if the reaction is unimolecular, the immediate products of 
the decomposition must react further since the stoichiomctry re- 
fers to two molecules of N0 2 per molecule of N 2 5 . Several possibil- 
ities exist for the initial decomposition: 


2 (3) 

The first can be ruled out because it involves a change in 
:he electron multiplicity in the initial step and thus is pre- 
jumably very slow. *he second has a '-> H which is endo thermic by 
>1 kcals.. E , however, is only 2l_i .6 kcals, and the -H for the rate- 
ietermininO step cannot exceed this. Reaction 3 has a oH of 
-l|.3 kcals. in good agreement with the experimental value and was thus 
tccepted as the initial step for quite some time. 

The low pressure investigation of 'this decomposition showed 
-hat it remained first order down to 0.05 mm» At lower pressures 

N 2 5 


N 2 3 


o 2 

N 2 5 


N 2 4 


N 2 5 




N0 2 




I - ■ 



L • 

. ■ 






• • . . . .••.■■ •■. 

- ■• ' ' ' - . 

■ . • . ; ■ J ■ 

■■ ' 





■ ... • ■:• - 





• • ■ i 



• : 



the apparent order with respect to N 2 5 increased to a value of 
1.8 at 10-3ram (10). This change in order is in accord with the 
Lindemann theory, except that it takes place at much too low a 
pressure. Using even the most favorable assumptions as to the 
molecular diameter and the number of degrees of freedom contribut- 
ing to the vibration of the molecule, calculations showed that the 
rate constant should have been appreciably changed at a pressure of 
0.1 mm. 

In 1947, Ogg showed that the experimental data could be ex- 
plained by the following mechanism (11): 

N 2 5 2 N0 2 + N0 3 (fast) k x and k 2 
N0 2 + N0 3 ■*■ N0 2 + NO + 2 (slow) k 3 

N0 3 + NO -> 2N0 2 (fast) K 4 

Using the steady state assumption for (N0 3 ) the rate of decomposi- 
tion of the pent oxide is 

-d(N 2 0s) *' = 2kik a (Ng0s) »' 2KiK 3 (N 2 05) 
dt k 2 + 2k 3 k 2 

This reaction 'is not unimolecular at all on this basis but involves 
a prior equilibrium presumably with a third body followed by a 
biraolecular rate -determining step. 

This mechanism preducts the falling -off of the rate constant 
at low pressure since the unimolecular dissociation of N20 5 into 
N0 2 and N0 3 is governed by activation through colision, and at very 
low pressures the rate will become equal to the rate of activation 
tby collision. 

Presumed proof of this mechanism has been offered by several 
investigators. Ogg has substantiated it by investigating the ex- 
change between N 2 15 5 and N0 2 . The rate constant for this reaction 
should be k x and should fall off as the total pressure is decreaded. 
This was observed experimentally (12). The reaction of NO with 
£N 2 5 should offer an alternative approach to determining ki since 
.the mechanism for this reaction contains ki as the slowest step. 
Mills and Johnston (13), Hisatsune, Crawford and Ogg (5), and 
Jach (14) have separately confirmed this. .. 

Further support for the Ogg mechanism comes from the spectro- 
scopic identification of N0 3 in the pent oxide -catalyzed decomposi- 
tion of ozone (15), and by the fact that the N0 3 Intermediate can 
explain quantitatively the oxidation of NO by 3 , the reaction of 
N2O5 with NO and the decomposition of ozone catalyzed by N 2 5 (lo). 
In addition, the shock tube experiments of Schott and Davidson have 
given direct measurement of the equilibrium between N 2 5 , N0 2 , and 
||N0 3 and have explained the subsequent decomposition of N0 3 (2). 

' ' - '■ ' - 


- 1 


• ■■■'. ■ 

i "■ •; 

[•■■■ : •■ 


. ""- ■ 


- . . • • 

• ■ • - ' , • ■ - 

- '- : ' ■ to ■ . ■ 


- " " ; 

'■ £ . 


• - • : - - ■■ ' 

: ■ . . . . • 

■ ■ • [TO09 

i . lo ; ■ .. .; - 

' ■ ■ ' I . . • : 



1. H. S. Johnston, Ann. Rev. Phys . Chem., 8, 263(1957). 

2. G. Schott and N. Davidson, J. Am. Chem. Soc . , 80, 1841(1958). 
p. P. Daniels and E. H. Johnston, J. Am. Chem. Soc, 43, 53(1921). 

4. F. Kaufman and J. R. Kelso, J. Chem. Phys., 23, 1702(1955). 

5. I. C. Hisatsune, B. Crawford, and R. A. Ogg, J. Am. Chem. Soc, 
7£, 4648(1957). 

6. M. Bodenstein, Z. physik. Chem., 100 , 68(1922). 

7. W. A. Rosser and H. Wise, J. Chem. Phys., 24, 493(1956). 

8. P. G. Ashmore and B. P. Levitt, Research (Corr.), £, S25(1956). 

9. For general discussion consult A. F. Trotraan -Dickenson, Gas 
Kinetics, Academic Press, Inc., New York, 1955; or A. A. Frost and 

. R. G. Pearson, Kinetics and Mechanism , John Wiley and Sons, 
Inc., New York, 1953. 

LO. H. C. Ramsperger and R. C. Tolman, Proc Natl. Acad. Sci., U.S. 
16, 6(1930). 

LI. R. A. Ogg, J. Chem. Phys., 15, 337(1947). 

L2. B. A. Ogg, J. Chem. Phys., 18, 573(1950). 

L3. R. L. Mills and H. S. Johnston, J. Am. Chem. Soc, 73, 938(1951). 

L4. J. Jach, Trans. Faraday Soc, 51, 41(1957). 

L5- E . J. Jones and 0. R. Wulf, J. Chem. 'Phys., 5, 873(1937). 

L6. H. S. Johnston, J. Am. Chem. Soc, 73, 4542 .'(1951). 




■ .] 



. • 

I • . ' • . 

... ' 








Dean W* Dickerhoof February Z\\, 1959 

I. Introduction 

The electron-pair bond is used as the basic model to explain 
the chemical combination of atoms into molecules. There are numer- 
ous compounds, however, in which the number of bonds required to 
connect all the atoms is greater than the number of valence electron- 
pairs available from the participating atoms. Many simple compounds 
of this type can be explained in terms of a "three-center bond" in 
which one electron pair binds three atoms together instead of two* 
In more complex compounds of this type, the three-center bond is 
generalized into molecular bonds through the method of molecular 
orbitals. This theory gives a satisfactory, qualitative pic- 
ture for the existence and stability of these compounds. 

II. The Three-center Bond 

Consider two atoms d and 2 ) of the same element and an atom ( 3 ) 
I of the same or a different element being positioned relatively near 
each other so that they must be considered to be bonded together* 
If there are just two valence electrons (assuming also that no atom 
has unshared pairs to donate) among the three atoms, the conditions— 
[for the formation of the three-center bond would be met. This bond- 
ling is'best described'by the linear combination of atomic orbitals 
'(the L.C.A.O. method), (l) 

The energy levels obtained from the linear combination o two 
suitable atomic orbitals into new orbitals is given by: 

E + = E + £^2 — i — E -^ , where E is the 

1 + S, -> 



the two orbitalsj since H 12 possesses a negative value, the lowest 
energy of the system is given by E + which would be the bonding or- 
bital . 

In the treatment of three orbitals (2,3), it is first assumed 
that S 12 is a constant and that H 12 =H 23 =H and that H 13 =G. By the 
variational method of calculation, the energy of this system can be 
given by: 


G/2 + [(G/2) 2 + 2H2] 1 /' 

Eo = H - G 

E_ = H + G/2 - [(G/2) 2 + 2H2] 1 / 2 

By comparing the relative values of E + , E , and E_ at H/G = O , 
G/H = O , and G/H — 1 , we arrive at one bonding level (E + ) and tv/o 
antibonding levels. There are only two electrons, however* so the 




bonding level (E + ) Is exactly filled, and thus one electron-pair 
binds all three atoms. At G/H - O, it is called an "open three- 
center bond" and at G/H = 1, it is called a "central three-center 

The orbitals combined by this method must be of the same rela- 
tive energy, of the same symmetry along the bond axis or plane, and 
orthoganal. Consider diborane with boron atoms B A , B 2 and bridge 
hydrogens H !l and H f,t '. The four orbitals obtained by combining the 
atomic orbitals (Lj.) are: 

Yl - (<*i + &z + s* + s")/2 

y z = (pi + p 2 + s' - s")/2 

'fz = (oi + az ~ s* - s")/2 

% = (Pi + P 2 - s* + s")/2 

where o- is the 2S orbital of boron, s is the IS orbital of the hy- 
drogens, and p is a suitable 2P orbital of boron perpendicular to 
the boron-boron bond axis. }'\ and 7*£ are bonding orbitalsj the 
latter tv/o are antibonding. It is nov/ possible to take linear com- 
binations of these orbitals. The two bonding orbitals are given by: 

x, = >: + y z / i2 

x 2 = yi - / 2 / "yr 

By algebraic substitution and rearrangement, it can be seen that Xj. 
and X 2 are composed of ( Oj. + pi) / \2 , (o" 2 + P2 ) / .2 , and the 
S orbitals of the hydrogens. These terms would also result if the 
boron orbitals are considered to be SP hybrids. Thus the linear 
combination of pure S or P or even'SP hybrids leads to stable elec- 
tronic configurations for diborane. This approach can be extended 
to the higher boron hydrides and the other electron deficient com- 
pounds • 

III. Metal Alkyls (5) 

^t has been found that metal alkyls such as Al(Me) 3 , Al(Et) 3 , 
Be(Me) 2 , and Ga(Et) 3 exist as dimers. (6) 

Compounds such as (Me 4 Pt) , (Me 3 ClPt) , and (Me 3 In) are tet- - 
rameric in the solid state (7,8). Since there are not enough con- 
ventional electron-pairs available in these molecules to form such 
polymers, bonding must be explained by the three-center bond (9*10). 
It is postulated that when an atom with an unoccupied bonding orbital 
(usually a metal) is combined with atoms or groups that do not have 
any unshared pairs (e.g. hydrogen or methyl groups) the atoms will 
group themselves so that all orbitals are at least partially occupied 
(ll). By forming molecular orbitals from available atomic orbitals, 
all bonding orbitals can be completely filled. The acute metal-carbon- 
metal angle measured is thought to be necessary for effective over- 
lap of the metal and carbon SP 3 orbitals. 


IV. The boron hydrides 

Bonding in the boron hydrides ( 12,13) is explained (2,3) in terms 
of open three-center bonds, central three-center bonds, and linear 
combinations of these to form higher molecular orbitals embracing the 
f rameuo rk of bor6n atoms. They are all fragments of an icosahedron 
or an octahedron* 

All hydrides of boron are members of two homologous series with 
the formulas B n H n+4 and B n H n . 6# Two new hydrides B 6 H 10 and B 9 Hi 5 
have recently been isolated T1J4.) and their structures determined (l5j 
16 )♦ The structures of the boron hydrides have been investigated 
mathematically by a topological approach. Predictions have been 
made concerning the existence of other possible hydrides (17)« 

V* Other Boron Polyhedra 

The atoms in elemental boron are arranged to form an icosahed- 
ral structure. There are three crystalline modifications (l8) con- 
taining 12, 50 and 108 atoms of boron per unit cell which means 
1, L(.'(plus 2 boron atoms), and 9 Icosahedra respectively in the unit 
cell. The bonding is explained (19) in terms of molecular orbitals 
consisting of 13 bonding orbitals. The boron atoms donate 21+. elec- 
trons? so two electrons must come from other icosahedra or boron atoms 
(20, 18) ♦ Boron also forms an icosahedral network in boron carbide 
B12C3 and again must receive two electrons from the carbon chain (19)» 

Boron forms an octahedron in the metal borides IvB 6 (21). Here there 
are seven bonding orbitals and eighteen valence electrons. Six elec- 
trons are used in normal single bonds to other octahedra so that two 
electrons again must come from the metal atom. The boron atoms in 
B4CI4, are arranged tetrahedrally with respect to each other? in — 
addition each boron atom is bonded to a chlorine atom (22). A com- 
pound with the double molecular weight, B 6 C1 6 , has also been reported 
and its structure determined (23 )• It consists of a boron network 
surrounded by chlorine atoms, but it is not a regular polygon. 

VI. Interstitial Phases 

R» E. Rundle ( 21+) has considered the bonding In metallic car- 
bides, nitrides, and oxides as "electron-deficient" bonding. The 
properties and structures of these phases appear to require octa- 
hedral bonding for the first row elements, viiich would mean that — 
there are not enough electrons for all electron pair bonds. The con- 
cept of the "half bond" has been advanced by Rundle. 





1. C,. A. Coulsort, "Valence", Oxford University Press, Amen House 
L6nd0n, 1956 • 

2. W, H. Eberhardt, B r y C e Crawford, and William N, Lipscomb, J. 
Chem. Phys., 22, 989 (195U). 

3. William N. Lipscomb, J* Phys, Chem,, 6J., 23 (1957). 
km H # C„ Longuet-Higgins, Quart. Revs,, 11, 121 (1957). 

5. G . E . Coates. ' "Organo-Metal lie Compounds", Methuen and C . Ltd., 
London (1956U 

6. A* W» Laubengayer and w # F. Gilliam, ^. Am, Chem, Soc, 6^ 

• lj.77 (1914-1). ' 

7. R. E . Rundle and J* H. Sturdivant, J. Am. Chem* Soc, 69, 1561 
(1914-7). . . 

8. E. L.. Amma and R. £. Rundle, J, Am. Chem, Soc, 80, I4.II4.I (1958). 
9i R. E . Rundle, J. Phys. Chem,, 61, k$ (1957). 

10. E # Cartmell and G, w. A, Fowles, "Valency and Molecular Struc- 
ture", Butterworths Scientific Publications, London, 1956, 
page 236, 

11.- R. E, Rundle, J. Chem. Phys,, 1J, 671 (191+9). 

12; William N . Lipscomb, J. Chem. Phys,, 22, 985 (195U. 

13; Reference 10, page 226 

1)4., W. V, Kotlensky and R, s chaeffer, J. Am, Chem, Soc., 80, i+517 
(1958). • """ 

15. F. L. Hirshfeld, K. Eriks, R. E. Dickerson, et, al,, J, Chem. 
Phys., 28, 56 (1958) ' 

16, R, E. Dickerson, P. J. Wheatley, et, al., J. Chem. Phys,, 27, 
200 (1957). 

17* R, E. Dickerson and W, N. Lipscomb, J, Chem. Phys., 27, 212 
(1957). ... 

18. J. L. Hoard, R, *E. Hughes, and D. E. Sands, J. Am. Chem. Soc., 
' 80, 4507 (195S). 

19. ^7 C. Lonquet-Higgins and M. de V, Roberts, Proc, Roy, Soc., 23CA V 
110 (1955). ... 

20. L. V. HcCarty, J. S„ Kasper, F^ H. H rn, et, al., J, Am„ Chem. 

• Soc, 80, 2592 (1958). 

21. H, C. TTonguet-Higgins and M. de V, Roberts, Proc, Roy, Soc., 
22l4A ,336 (19514-) . 

22. Masao Atoji and W. N. Lipscomb, J . Chem. Phys., 2J,, 172 (1953). 

23. W. N. Lipscomb and R, A. Jacob son, J, Am, Chem, Soc, 80, 5571 
(1958). "" 

2I4., R. E. Rundle, Acta Cryst., 1, 180 (19I4.8). 




' . ■ . ' 








James E. Huheey March 3, 1959 


A» Preparation and Structure 

1« Diboron Tetrachloride was first prepared in 1925 
by striking an arc between zinc electrodes in the presence of BCI 3 (l). 
However, because of the low yields and handling difficulties almost 
25 years passed before further work appeared. Schlesinger and co- 
workers prepared B 2 C1 4 by passing boron trichloride vapors at low 
pressures through an alternating current glow discharge between 
mercury electrodes(2,3 ) . Holliday and Massey found that yields are 
improved in the glow discharge method if a direct current is used 
in place of alternating current(5)« A method utilizing microwave 
excitation has also been developed (14.) • 

From the infra-red and Raman spectra of gaseous and liquid 
B 2 C1 4 , VVartik and co-workers concluded that the molecule has the 
symmetry of a non-planar ethylene model(6). Lipscomb and co-workers 
found by an X-ray diffraction study that the molecule is planar in 
the solid, with the B-B bond stretched somewhat(7)» This difference 
in structure as determined by these tv/o methods could either reflect 
an error in the spectroscopic interpretation or the presence of 
sufficient lattice energy in the solid to twist the non-planar mole- 
cule into planarity upon crystallization, 

2 # Tetraboron tetrachloride is formed in small 
quantities as a by-product in the preparation of diboron tetrachloride 
(8). It appears as a yellowish crystalline residue when B 2 C1 4 is 
volatilized. It decomposes in aqueous basic solution with the liber- 
ation of six equivalents of hydrogen per mole. This would indicate 
six boron to boron bonds and suggests a structure formed of a 
tetrahedron of boron atoms (8). X-ray studies have shown that this 
picture is correct and that the chlorine atoms form a tetrahedron 
around and beyond the tetrahedron comprising the four boron atoms(9)« 

3» Octaboron octachloride is also obtained as a 
by-product in the preparation of diboron tetrachloride. The mole- 
cule consists of an irregular polyhedron with one boron and one 
chlorine atom at each corner(lO). 

B» Reactions of Diboron Tetrachloride 

1. Addi tion React ions -Pi boron tetrachloride forms 1:1 
addition compounds with hydrogen sulfide(ll) and ethers(3) and 1*2 
adducts with ethers(3)> thioethers( 11 ), phosphine( 1 1 ) and tertiary 
amines(3). The product of the reaction between diboron tetrachloride 
and trimethyl amine is especially interesting because molecular 
weight determinations indicate that no structure has as yet been 
proposed for the addition compound* 

2» Replacement Reactions -The chlorine may be replaced 
easily by various groups through solvolytic processes. '-he reactions 
probably proceed via the addition of a base to the Lewis Acid, B 2 C1 4 , 
followed by loss of chloride ion or hydrogen chloride. Thus 


■ t 







, • 


•■ - ' 





3 :-. 






• J 




• ■ 

■ ■' 




■ ■ ■ 
■ ' 



. - 

' -' - 
t 9 ■ 





hypoboric acid, B 2 (OH) 4 , or its esters may be produced by reaction of 
B 2 C1 4 with water or alcohols, respectively (2,3) • Likewise, N-sub- 
stituted amides have been prepared from B 2 C1 4 and secondary amines(3), 
Reaction with ammonia gives a product of composition (BNH) , but 
little is known about its structures (3 ) • 

At temperatures below ~78°C, hydrogen sulfide forms a normal 
addition compound with B 2 C1 4 . Above this temperature, further 
reaction occurs with the formation of BC1 2 SH, B 3 S 3 C1 3 , BC1 3 , and 
other products( 1 1 ) • Trichloroborsulf ole is of special interest be- 
cause of its ring s tructure( 12) • 

♦ 3. Oxidation Reactions -Pi boron tetrachloride reacts 
with chlorine, bromine, and oxygen to give BC1 3 and B X 3 (n=oxid? 
tion state of X). Mixed halides could not be isolatea. No reaction 
occurs when B 2 C1 4 is treated with iodine, sulfur, or white phosphorus 
(13)* The reaction with cyanogen is complex, yielding a product 
whose composition corresponds to the empirical formula B 2 C1 4 - 1 e5(CN) • 
When heated, BC1 3 is released* Wartik has proposed structural 
mulas to explain this behavior( ll|) • 


Lower iodides have been prepared( 15) . Treatment of ' 
with a radio frequency discharge at low pressures yields di boron 
tetraiodide. The compounds B I (x^y) and (Bl) n have also b 
reported. ^ 

Diboron tetrabromide may be obtained in small amounts^ 
together with a brown-black amorphous powder of variable composition, 
by glow discharge In a mixture of Ar and BBr 3 (l6)» 

Diboron tetraf luoride has been prepared by treating B 2 Cl 4 
with antimony trifluoride. It behaves similarly to the chloride, 
forming etherates and a tetrameric addition compound with trimethyl- 
amine(l7). Its structure has been shown by an X-ray study to be 
planarC 18) . 


. ■ - 


' I 

• . 


1. A. Stock, A. Brandt, and H. Fischer, Ber., 5j3, 6i;3(l925). 

2. T. Wartik, R. Moore, and H. 1 • Schlesinqer, J. Am. Chem. Soc., 

71, 3265(1914-9). 

3. G # Urry, T # V/artik, R. E. Moore, and H. I. Schlesinaer, J. Am. 

Chem. Soc., J&* 5293(1951+). 
U. J. w . Frazer and R. T. Holzman, J. Am. Chem. Soc., 80, 2907 

5. A. K. Hoi li day and A* G. Massey, J. Am. Chem. Soc, 80, 147-4 


6. M. J. 

A m 

7. M. At 





• Linevsky, E; R. Schull, D. E. Mann, and T. Wartik, J. 
A m . Chem. Soc., 7£, 3287 (1953). 

toji, W, Lipscomb, and P. J. Wheatley, J. Chem. Phys. 23, 
1176 (1955)? 27, 196 (1957). 

16. A; Pfugmacher and 'W # Diener, Angew. Chem., 69, 777 (1957 

17. A # Finch and H. I. Schlesinger, J. Am. Chem» Soc., 80, 3573 


18. L. Trefonas and W # N. Lipscomb, J. Chem. Phys., 28, Sh (1958). 



i 4 


L. J. Sacks March 10, 1959 

I. Introduction 

A base may be defined either as a proton acceptor or from a more 
general view as an electron pair donor [Lewis (1)]. A Lewis acid, 
accordingly, is an electron pair acceptor. Many reactions may 
therefore be classified as acid-base processes. 

The amines offer several advantages in the study of basicity: 
(a) They offer a wide range of size and configuration; (b) they 
are soluble in both polar and non-polar solvents; and (c) many undergo 
reactions in the gas phase. 

II. Measurement of Basicity 

While the definition of a Lewis base as an electron pair donor 
offers a definite criterion for basicity, a method for evaluating 
this characteristic is less universally accepted. Measurements of 
pKb in aqueous solution indicate (Table 1) that, in general, primary 
amines are stronger bases than ammonia, secondary amines are the 
strongest bases but tertiary amines weakest (2-4). One would expect, 
however, that the third replacement of hydrogen by an alkyl group 
should continue the monotonic increase in basicity. Further 
irregularities are found with more complicated amines ( vide infra ) . 
Yet it would seem that some relationship between basicity and 
substituent groups must exist. The remarkable success of the Hammett 

and Taft equations (5): log k/k = cv or T*/? * ±n correlating data 

for thousands of organic reactions led to attempts at similar 
correlation of the basicity of amines. Hall found that a plot of 

pKa for substituted ammonium ions vs. _ ** [iT* values, the .. "polar 

substituent constants" determined by Taft, are additive properties 
of the various substituent groups (5)] gave essentially a linear 
relationship for tertiary amines and somewhat poorer correlations for 
primary and secondary amines (2). Brown and Okamoto (6) modified the 
Taft equation to account for charged species as well as neutral 
molecules. Edwards (7) correlated data for twelve highly diverse 
manifestations of basicity, including complex ion formation, displace- 
ments on carbon, substitution in cis -[Coen 2 N0 2 H 2 3 +2 and even 
quenching of fluorescence by means of the equation: 

log *L_ - qe + 0H, where k is the equilibrium or rate constant, E n 

is a "nucleophilic constant" characteristic of the donor, H is the 
relative basicity of the donor toward protons and a and £ are 
substrate constants. Other criteria for basicity include solubiliza- 
tion of silver oxide (8), dipole moments (9), catalytic activity 
(10-12) and equilibrium constants as estimated from spectrophoto- 
metry (10,15) or potentiometric (5) methods. 


III. The Fundamental Manifestation of Basicity 

The fundamental manifestation of basicity is the energy released 
when an acid and base combine. Generally, this energy is reported as 
AH, and gives a quantitative measure of the stability of the bond 
[see e.g., (14)]. As pointed out by Brown (14) and others (3,4), the 
medium in which the reaction is carried out will have a decided effect 
on the heat of reaction because of interactions with reactants and/or 
products. Hall (2,3,15) and Pearson (4,10) discuss the effects of 
hydrogen bonding of primary and secondary amines and ammonia in 
protonic solvents. Apparent irregularities in the order of basicity 
from that expected from polar effects of substituents are explained 
on the basis of interactions of amine and adduct with solvent. 
Differences between predicted and experimental orders of basicity 
may also be explained in terms of various "strains" that the amine 
undergoes in forming the adduct (16). 

IV. Solvation Effects 

The solvation of an ammonium or substituted ammonium ion involves 
not only gross solvation of the charged ion but also specific inter- 
actions between the hydrogens and the solvent molecules. Pearson and 
Vogelsong (10) cite evidence that in hydration the latter effect 
amounts to ca. 8 kcal per mole per hydrogen. Altshuller (17) has 
estimated that for ammonium ion hydration the entropy decrease is ca. 
7 e.u. above that for a monatomic ion of comparable size. This is 
attributed both to charge distribution over the four hydrogens and 
to hydrogen bonding with water molecules. Hall claims that the 
energy of hydration of free amine bases is small and essentially 
constant (3), a rather surprising conclusion, considering the 
diversity of hydrogen bonding possibilities. 

Attempts to eliminate solvation effects by employing non-polar 
solvents are useless. A pointed out by Pearson (3), benzene can 
exert an appreciable interaction with ionic species. Even alkanes 
retain a dielectric constant of ca. 2 (3), hence can interact 
appreciably with ionic solutes, especially the proton. The atraction 
of the proton for non-positive molecules has been substantiated by 
quantaum mechanical calculations (18). The differences between the 
various classes of amines are thus explained (3,10) in terms of the 
degree of solvation of the corresponding ammonium ions. This 
predicts an order of basicity: NH 3 RNH 2 R2NH R3N, and accounts 
for the low basicity of tertiary amines. Combined with polar effects, 
this explains amine basicity fairly well for many cases. However, 
solvation effects on the acid and adduct also change from solvent 
to solvent, but are neglected here. Obviously, the nature of the 
medium will also have a profound effect on the acidity of the acceptor. 
This effect is seen in the measurements by Hall in non-aqueous 
solvents (15). 

V. Elimination of Solvation Effects 

To eliminate the permuting effects in solution, Brown, et al . , 
measured a wide variety of amines in the gas phase, using BMe 3 as 
reference acid (19-24). A summary of their results is presented in 
Table 2. To interpret these and related results, Brown proposes 










• ' 5 ■ 










three types of steric strain (16): (a) F-strain, associated with 
bulky groups at the faces of the acid and/or base; (b) B-strain, 
resulting from compression and bond-bending within the groups 
attached to an atom, accompanying the formation of a bond; and (c) 
I-strain, an increase in the internal strain of a cyclic structure 
resulting from a new bond causing a change in bond angles. 

These strains rationalize most — but not all — of the gas 
phase data and are applicable to solution phase (4,14). However, 
Hall says (2) that for tertiary amines the plot of pKb vs. 

]T q-* shows no deviations as would be expected if B-strain were 

operative. Independent evidence for steric strain is found in the 
NMR spectra of sterically hindered dihydrophenanthrenes (25). 

Brown's measurements of AH cannot be compared with equilibrium 
values, which must have variable temperature coefficients and hence 
cannot possibly be considered as fundamental manifestations of 
basicity. Hence, while certain of the arguments proposed in favor 
of solution effects are quite reasonable, they cannot be evaluated 
by any of the data which Hall or Pearson and Vogelsong present. 

Brown's use of AH as the measure of bond energy seems 
unfortunate. The actual energy is AE, and the difference, 
typically of the order of 800 cal per mole, is not a negligible 
quantity. This will be especially pertinent if heats in solution 
(14) and in the gas phase are to be compared. Also, as shown by 
Tamres (26), certain corrections must be made to Brown's gas phase 
values to corrent an experimental error. 







Table 1. pKb for Some Simpler Amines" 



NH 3 



n-BuNH 2 

Me 2 NH 

(n-Bu) 2 NH 

Me 3 N 



(n-Bu) 3 N 

Et 2 NH 

n-C 6 NH 2 

Et 3 N 

1. Value for ammonia from Willard, H.H. and Furman, NH. , Elementary 
Quantitative Analysis, 3rd Ed., p. 122, Van Nostrand Co. , (1940) , 

Other values 

from re 



* ■> 

r -K- * * 



Amine -BMe 3 Compounds; 

Gas Phase 





kcal mc 


K D 

AS Ref. 

NH 3 



39.9 20 

CH 3 NH 2 



40.6 20 

(CH 3 ) 2 NH 
(CH 3 ) 3 N 



43.6 20 



45.7 20 

(CH 3 ) 3 N 



45.9 21 




40.1 22 




44.3 22 




43.5 22 




45.0 22 

Et 2 NH 



44.1 22 

EtNH 2 



43.0 24 








dissociated to 1 

neasure 19 









Et 3 N 



dissociated to i 

neasure 23 

n-C 3 -NH 2 



43.0 24 

n-C 4 -NH 2 



43.2 24 

n-C 5 -NH 2 



43.9 24 

n-C 6 -NH 2 



43.2 24 












. . '. 




■ .. : 



1. Lewis, G.N., Valence and the Structure of Atoms and Molecules , 
p. 142, The Chem. Catalog Co. (1926) . 

2. Hall, H.K., Jr., J. Am. Chem. Soc, 79, 5441 (1957). 

3. op. cit ., 5444. 

4. Pearson, R.G. and Vogelsong, D.C., ibid., 80, 1048 (I958). 

5. Taft, R.W., Jr., in Steric Effects in Organic Chemistry , Ch. 13, 
M.S. Newman, Ed., Wiley (I956). 

6. Brown, H.C. and Okamoto, T., J. Am. Chem. Soc, 80, 4979 (1958). 

7. Edwards, J. 0., ibid., 78, I8I9 (1956); 76, 1540 (1954). 

8. Britton, H.T.S. and Williams, W.G., J. Chem. Soc. (1935), 796. 

9. Bax, CM., et al . , ibid ., (1958) 1258. 

10. Pearson, R.G. and Vogelsong, D.C., J. Am. Chem. Soc, 80, 1038 

11. Gerrard, W. and Lappert, M.F., Chem. Rev., 58, 1082, 1101-4 
(1958). ~ 

12. Edwards, J.O., J. Am. Chem. Soc, 76, 1540 (1954). 

13. Kilpatrick, M. and Hyman, H.H., ibid ., 80, 77 (1958). 

14. Brown, H.C. and Holmes, R.R., ibid ., 78, 2173 (I956). 

15. Hall, H.K., Jr., J. Phys . Chem., 60, 63 (1956). 

16. Brown, H.C, Rec Chem. Progr., 14, 83 (1953). 

17. Altshuller, A. P., J. Am. Chem. Soc, 77, 3480 (1955). 

18. Longuet-Higgins, H.C, Rec. Trav. Chim. , 75, 825 (1956). 

19. Brown, H.C. and Barbaras, G.K., J.Am. Chem. Soc, 6<2, 1137 (1947). 

20. Brown, H.C, et al., ibid ., 72, 2923 (1950). 

21. Brown, H.C and Gestein, M., ibid ., 72 , 2923 (1950). 

22. op_. cit ., 2926. 

23. Brown, H.C. and Taylor, M.D., ibid ., 69, 1332 (1947). 

24. Brown, H.C, Taylor, M.D. and Sujishi, S., ibid ., 73, 2463 (1951). 

25. Reid, C, Mol. Spectroscopy,!., 18 (1957) as reported in C.A., 51, 
11849a (1957). 

26. Tamres, M., Private communication of unpublished results. 

I ■ 




D. K. Straub March 17, 1959 


The phenomenon of araphoterlsm probably occurs In all solvent 
systems, I.e. in both protonic and aprotonlc solvents. In the pro- 
tonic solvent water, aluminum, zinc and chromium(lll) form complex 
hydroxides . Sulf ito complexes have been shown to form in the apro- 
tonlc solvent sulfur dioxide (l). Schmitz-Dumont has been investi- 
gating the formation of complex amides in liquid ammonia. It is 
reasonable to assume that the formation of these complexes takes 
place in a stepwise fashion, even though few of the probable inter- 
mediates have been isolated. Deammonation of complex amides takes 
place in liquid ammonia much more readily than dehydration of complex 
hydroxides in water. 

Transition metal amides 

Preparation of the transition metal amides and complex amides 
entails the reaction of a soluble anhydrous salt of the metal with 
potassium amide in liquid ammonia. Probably a complex ammine is 
formed first upon solution of the metal salt in liquid ammonia; 
protons are then removed stepwise by the strong base, amide ion; 
deammonation may take place (depending on temperature) to give the 
compounds that are actually isolated: 

M x -> M(NH 3 )* * M(NH 2 )^(NH 3 )^ * amide 
y z y— z 

amide -*■ mixed amide-imide ■+■ imide ■* mixed imide-nitride -> nitride 

In many cases, spontaneous deammonation occurs, so that the inter- 
mediates have not been isolated. The amide compounds obtained from 
transition metal salts may be generally characterized as pyrophoric, 
polymeric, oxidizable, and easily hydrolyzable substances. The amides 
and complex amides are readily attacked by the acidic ammonium ion; 
the deammonated compounds are less readily attacked. 

Titanium . Greenish-black titanium(lll) amide reacts with excess po- 
tassium amide to form the green imide, K[Ti(NH) 2 ] (2). The corre- 
sponding tltanyl amide, TiO(NH 2 ) 2 , is assumed to contain bridging 
oxo and ami do groups, and reacts with excess amide to give the imide 
TiO(NHK) 2 (3). Upon heating, (TiO) 3 N 2 and (TiO) 3 N 4 Ka are formed 
from titanyl amide and imide, respectively. Both these nitrides are 
decomposed to titanium(ll) oxide when strongly heated. 

Zirconium . No simple amide of zirconium has been obtained. The 
rather complex ammonolysis product resulting from the action of am- 
monia on zirconium tetrabromide reacts with amide to form a ye&low 
substance whose composition corresponds to Zr(NK) 2 «2NH 3 (4). 

'- : ' •'-. ; 




■■■ 3 

■ ■•-.. 





'■ : . blm .'.■■' 

: • . 
; • ■ : ■ : 

, ■• 8 I 

tup- . . 

■ : . 


." •' 



' ■ . ■ 

.■ I 

■ ' ' ' • - 

■ . ' 



Vanadium * The black imide, KV(NH)2 results from potassium amide and 
K 3 LV(SCN) 6 ] (5). Evidently deammonation of the hypothetical amide 
takes place so rapidly that only the imide can be isolated. This 
imide reacts slowly with ammonium ion in liquid ajmmonia; thus It is 
impossible to prepare the imide -amide, HNVNH 2 , in accordance with the 
following equation: 

V(NH) 2 K + NH 4 + * HNVNH 2 + K + + NH 3 . 

Tantalum . Ammonolyzed tantalum pentabromide yields Ta(NH 2 ) 4 NH.K and 
Ta(NH 2 ) 3 (NHK) 2 with excess potassium amide (6). 

Chromium. Light pink Cr(NH 2 ) 3 is obtained from potassium asiioe -2nd 
TcTTNHaTe ] ( NO3 ) 3 in liquid ammonia (7)« With excess amide, 
[Cr(NH 2 j 4 ]nKn results. This splits off ammonia to give viol : black 
[Cr(N^gf| c ]nKn. 

With chromium complexes in liquid ammonia, the amraonolysis re- 

[Cr(NH 3 )e] +3 + NH 3 ^ [Cr(NH 3 ) 5 NH 2 ] +2 + NH 4 + 

[Cr(NH 3 ) 5 NH 2 ] +2 + NH 3 $ [Cr(NH 3 ) 4 (NH 2 ) 2 ] +1 + NH 4 + etc., 

lie far to the left (8) . However, the equilibrium can be shifted by 
removal of the NH 4 with a strongly protophilic anion, like amide or 
an alcoholate. Thus [Cr(NH3) 6 ] (0CH 2 C 6 H5)3, prepared from the complex 
nitrate, potassium amide, and benzyl alcohol, reacts with ammonia to 
give bright gray-violet 'amidochroraium dibenzylate", [H 2 NCr(OCH 2 C Q H 5 ) 2 ] 
(after drying in a stream of nitrogen) . When freshly prepared, and 
before drying, this compound contains coordinated ammonia; it then 
will react with one mole of potassium amide to form violet 
HNCr 2 (NH 2 ) 3 (OR) 2 K, and with five moles to form coffee brown Cr(NH) 2 K. 
Amidochromium dibenzylate is assumed to possess bridging amido and 
benzylate groups. It also reacts with diphenylketene to give brown 
diphenylacetylamidochromium dibenzylate, (OR) 2 CrNHCOCH(C e Hs) 2 , and 
with phenyl isocyanate to give what may be a biuret-like compound, 
containing ©ne benzylate and two isocyanate groups per chromium. 
These polymers are reversibly depolymerized in pyridine sblution. 

If the benzylate anion is replaced by the phenolate anion, no 
ammonolysis takes place, due to the higher acidity of the phenol and 
consequent lower basicity of the anion (9). H 2 NCr(OCeH 5 ) 2 results 
when the phenylate is heated. 

Molybdenum and tungsten . With excess amide, ammonolyzed molybdenum 
pentachloride and tungsten pentabromide give compounds having the 
composition Mo(or W)(NK) 2 NH 2 (10) . These have not yet been obtained 

Manganese . Manganese (11) thiocyanate gives manganese (11) amide with 
potassium amide; this dissolves in excess amide, yielding light 
yellow Mh(NHK) 2 «2NH 3 (11). 

. ■ . 

• ■ 

■ I 

■ ' I 

. : 


I •■■■■ 


l e l LV 

■ . 

1 ! . ■ 

' ■ 

■ ' a 
■ ■ 

. . 





rnoo .- ■ ; 


■ ■■- ' ■■ 


nq it ■■■ i ... a " ■ ■ ■'• i 

' ■ ■ ■■ .■■■■ ' - 

'■' •' ■ ' • ■ - is "mi •. ' . . • ' =• ■ . .' . 

tf; • i v ? .. ■ 

•'" • ■ ■ " • ' ' 

drill -. . "•■ .- . . . • 

' ■ tjiibj as ■ . ■ ■:. \ .,■<■■■ -\ ■'■■ ■■ 

■ '•(;■ •• ■ ; ' ■ ' ■ 

•■:■. COJ ■ ' . i 6390X9 1 . 

' '"• I ■ -■■' • ' . ■■' ' ■ '. 


/ .. 



Cobalt . At low temperatures, violet crystals of Co 2 (NH 2 ) 3 (NHK) 3 
can be Isolated from the reaction between brown cobalt (111; amide 
(from hexamminecobalt(lll) nitrate and potassium amide) and excess 
amide; at higher temperatures, the double nitride of potassium and 
cobalt, C02N3K3, is obtained (12, 13). 

Nickel. Nickel (11) amide, a red crystalline subntance, will dis- 
solve in excess amide (14) . From the resulting solution, red crystals 
of the composition 2Ni(NH 2 ) 2 -5KNH 2 are obtained. 

Thorium. When four moles of potassium amide are allowed to react 
with one mole of K 2 [Th(N03) 6 J (chosen because it is an easily pre- 
pared, anhydrous, soluble salt of thorium) white HNTh(NH 2 ) 2 results, 
rather than the amide (15) • At 50°, this product loses ammonia, 
forming Th 2 (NH) 3 (NH 2 ) 2 ; at 100°, Th(NH) 2 , and at 130°, yellow TI13N4, 
are formed. Colorless crystals of octamminethorium (IV) iodide, 
[Th(NH3) 8 ]l4j result from the reaction between thorium imide diamide 
and ammonium iodide. A variety of products are formed with excess 
amide: (HN) 2 ThNH 2 K, HNTh(NH 2 ) 3 K, and Th(NH) 3 (NH 2 ) 4 K 2 , depending upon 
the amount of excess potassium amide used. All of these are white 
and explode in air. These complexes decompose with rising tempera- 
ture to Th 2 (NH) 5 K 2 , K 3 Th 3 N 5 , and finally ThN. 

Uranium . Uranyl amide, from potassium amide and K[U0 2 (N0 3 )33 Is a 
brown substance which gives a diuranate, rather than the expected 
hydroxide, when hydrolyzed: 

2U0 2 (NH 2 ) 2 + 3H 2 -» (NH 4 )2U 2 T + 2NH 3 

Its structure probably contains bridging araido groups. Uranyl amide 
reacts with ammonium iodide to form the diammine iodide, 
U0 2 (NH 3 ) 2 I 2 , and with potassium amide to form the imide U0 2 (NH) 2 K 2 . 
Upon heating with ammonia in an autoclave, uranyl amide forms the 
imide U0 2 NH, and finally uranium(lV) oxide. (U0 2 ) 3 N4K e results from 
heating U0 2 (NH) 2 K 2 . 

: ~- 




1. L. F. Audrieth and J. Kleinberg, Non -A queous Solvents , ■.-,' 
John Wiley and Sons, Inc., New York, 19557" 

2. 0. Schmitz-Dumont, P. Simons, and G. Broja, Z. anorg. Chem., 
258 , 307 (1949). 

3. 0. Schmitz-Dumont and F. Ftfchtenbusch, Z. anorg. Chem., 284 , 
278 (1956). 

4. E. W. Bowerman and W. C. Fernelius, J. Am. Chem. Soc, 6l, 

121 (1939). , , o> 

5. 0. Schmitz-Dumont and G. Broja, Z. anorg. Chem., 255 * 299 (1948). 

6. R. Levine and W. C. Fernelius, Chem. Revs., ^4, 449 (1954) . 

7. 0. Schmitz-Dumont, J. Pilzecker, and H. F. Piepenbrink, Z. anorg, 
Chem., 248 , 175 (1941 ). 

8. 0. Schmitz-Dumont and H. Schulte, Z. anorg. Chem., 282, 253 


9. 0. Schmitz-Dumont and R. Fischer, Z. anorg. Chem., 285* 303 

10. F. W. Bergstrom, J. Am. Chem. Soc, 47, 2317 (1925). 

11. F. W. Bergstrom, J. Am. Chem. Soc, Jb, 154f> (1924). 

12. 0. Schmitz-Dumont and N. Kron, Z. anorg. Chem., 280, 180 (1955). 

13. 0. Schmitz-Dumont, J. Pilzecker, and H. F. Piepenbrink, Z. anorg. 
Chem., 248 , 175 (194l). 

14. G. S. Bohart, J. Phys. Chem., 1£> 537 (1915). 

15. 0. Schmitz-Dumont and F. Raabe, Z. anorg. Chem., 277 * 297 (1954). 

16. 0. Schmitz-Dumont, F. Ftfchtenbusch and H. Schneiders, Z. anorg. 
Chem., 2H., 315 (1954). 




Donald A. V/enz March 2l\. 9 1959 

I. Introduction. 

Interest has been manifested recently in compounds containing 
phosphorus bonded directly to silicon. Emeleus and coworkers (l) 
have prepared a series of iodo-si lyl-phosphines from white phos- 
phorus with iodosilane, Iodosilane also reacts with trialkyl- 
phosphines to form the phosphonium type compounds R 3 P»SiH 3 I. 

Fritz observed that si lane, on heating above I4.OO , gives 
hydrogen atoms and silyl radicals (2). This finding suggested 
that si lane should react with ethylene, which was found to yield 
a number of products. It was therefore postulated that si lane 
should undergo a similar reaction with other hydrogen compounds, 
including phosphine. 

II, Preparation of si lylphosphine. 

Si lylphosphine was prepared by passing'a mixture of equal 
parts of si lane and phosphine through a 60 era. reaction zone at 
5>C0° (3,1+) • The effluent gases were condensed in traps cooled with 
liquid nitrogen. The hydrogen produced was pumped off, and the 
condensed products and reactants were separated by fractional 
distillation, A product was isolated that had a molecular weight 
of 6I4. and contained silicon and phosphorus in a ratio of 1:1, 
corresponding to the formula SiH 3 PH 2 » Extrapolation of a linear 
plot of log p versus 1/T gave a value of 12.7° at 76O mm, for 
the boiling point of s ilylphosphine. 

III* Reactions (I^,5)» 

1. With hydrogen bromide. 

If HBr is passed into a closed system containing SiH 3 PH 2 at 
-80 , bromosilane and phosphine are formed. Phosphine was proved 
to be present by its reaction with mercuric chloride to form a 
yellow precipitate of P(HgCl) 3 , Cleavage of the silicon-phosphorus 
bond by hydrogen bromide instead of formation of an addition com- 
pound may seem surprising, but Stock and Somieski (6) had observed 
analogous cleavage of the nitrogen-silicon bond In (SiH 3 ) 3 N by 
hydrogen chloride. 

2. Hydrolysis. 

a." Alkaline, Si lylphosphine is not hydrolyzed readily 
by cold water. However, if dilute sodium hydroxide or ammonium 
hydroxide is added to condensed si lylphosphine in the presence of 
methanol at liquid nitrogen temperatures and the mixture is then 
allowed to warm until it melts, hydrogen, phosphine, small amounts 
of silane, and silicic acid are formed. By observing the volumes 
of gases formed, it was found that three moles of hydrogen are 
produced per mole of si lylphosphine, upon hydrolysis. Hydrogen 
arises from hydrolysis of the Si-H bonds after the molecule is 


b. Acid. In the presence of methanol at -50°, silyl- 
phosphine is hydrolyzed by dilute hydrochloric acid to give 
hydrogen, phosphine, small amounts of silane, and a solid silicon 
oxyhydride. Apparently the molecule is again split at the Si-P 
bond, and the postulated intermediate, H 3 SiOH, is presumed to 
dehydrate to siloxane, which disproportionates to hydrogen and 
silicon oxyhydride (7). T he latter compound may be cf the type 
( SiH 2 0) . To substantiate this, chlorosilane was subjected to 
the same x reaction conditions; the same silicon compounds were 

3. Alcoholysis. 

a. By ethanol and sodium ethoxide. Si lylphosphine is 
not alcoholyzed by ethanol at -80°,.but gas evolution is observed 
when ethanol and sodium ethoxide are used* Hydrogen and phosphine 
gases are evolved in the ratio 3:1, just as in the aqueous alka- 
line hydrolysis, Alcoholysis appears to proceed through cleavage 
of the molecule and subsequent reaction of the resulting ester 
with ethanol, 

b. By ethanol and hydrogen chloride. When ethanol, 
saturated with hydrogen chloride gas, reacts with SiH 3 PH 2 at 

-80°, there is slow evolution of phosphine; small amounts of hydro- 
gen and silane are also formed. When gas evolution ceases and the 
solution is warmed to room temperature, a vigorous evolution of 
hydrogen, with traces of silane and phosphine, is observed. The 
reaction at -80° is believed to entail cleavage of the Si-P bond 
by ethanol, forming H 3 SiOC 2 H 5 and phosphine. The second reaction 
results in the decomposition of the ester into hydrogen, traces 
of silane, and a polysilicic acid ester that still contains 
silicon-hydrogen bonds, 

L(., Ammonolysis. 

Silylphosphine is decomposed by liquid ammonia, with forma- 
tion of phosphine, silane, and solid Si-N-H compounds as reaction 
products. The silicon-nitrogen-hydrogen compound is thought to 
be similar to the product obtained by Stock and Somieski (6), 
having a formula (SiH 2 NH) . This product is assumed to arise 
from cleavage of silylphosphine by ammonia to form phosphine and 
H 3 SiNH 2 j the latter then undergoes further decomposition. 

5. Thermal decomposition. 

On heating silylphosphine in a glass system in which the 
pressure can be measured, the variation in pressure v/i th increasing 
temperature becomes non-linear at about IjOOo, indicating decomp-^ 




silylphosphine itself, 

6. With ethylene. 

Since silane and silylphosphine under go t hernial decomposition 
by similar routes, it was expected that SiH 3 PH 2 would also react 
with ethylene. Pressure measurements on a mixture of silylphosphine 
and ethylene at 1+30° indicated that such reaction does take place. 


The products are difficultly volatile, oily liquids wi th strong 
reducing characteristics. Quantitatively, these products com- 
prise silicon-organic compounds in which phosphorus is bonded 
directly to silicon. 


1* B. J. Ayiett, H. J # Emeleus, and A.G. Maddock, J. Inorg. Nuclear 
Chem., & 187 (1955). 

2. G. Fritz, Z, Naturf orsch. , 7b, 207 (1952). 

3. G # Fritz , ibid, 8b, 776 (1953). 

1+. G. Fritz, Z. anorg. allgem. Chem., 280, 332 (1955). 

5. G. Fritz and H # o. Berkenhoff, ibid, 289, 250 (1957). 

6. A. Stock and C. Somieski, Ber., 5]±, 7^0 (1921). 

7. A. Stock and C, Somieski, ibid, £6, 132 (1923). 


. . 

■ I 



> ■ ', 

- ■ I 

* » 

IT • J * *r >\ r 

■ r 

* *• .. ? 

■ '■■'■■,.' t 

4 ' 4 - 

! ' 


Some Inorganic Ring Systems 

Richard L. Carlin March 31, 1959 

There are many inorganic compounds and groups whose structures 
are depicted as ring systems. Since a rather wide range of elements 
comprise such structures, there is a corresponding wide range in the 
properties and chemistries of these substances. Fundamental trends 
covering the following ring systems will be discussed: borazene, 
s-triazine, the phosphonitrilic halides, and tetrasulfur tetranitride. 

I. Chemical Properties 

A . Preparation 

While the mechanisms of ring formation are quite obscure, there 
are several well -known synthetic methods whereby their synthesis can 
be accomplished. Borazene, B3H3N3H3, has been made by the pyrolysis 
of the ammoniates of the boron hydrides, and more recently by the re- 
action of a borohydride with ammonium ion (l) and by the reduction of 
B3CI3N3H3 (2); derivatives ai*e made by somewhat similar reactions 
(3-6). The more-recently discovered s-triazine, (HCN)3, has been made 
by the acid-catalyzed polymerization of HCN (7,8) and by the reduction 
of the methylmercapto triazine (9). The trimeric phosphonitrilic 
halides, (NPXa)3, are best made by the reaction of the phosphorus 
pentahalide with the ammonium halide (10-13); the fluoride is made by 
treating the chloride with KSQ*F(l4). Tetrasulfur tetranitride, N 4 S 4 , 
has been prepared by the reaction of sulfur with liquid ammonia; the 
preferred method entails interaction of ammonia with a sulfur chloride 
(11, 15-17). 

B. Structure and Physical Properties 

All these compounds and their derivatives, except N4S4, are 
characterized as six-membered planar cyclic structures, with alter- 
nating Z-N atoms (18-23). Tetrameric (NPC1 2 ) 4 (24) and S 4 (NH) 4 (25) 
represent puckered ring systems, while N 4 S 4 consists of a bisphenoid 
of sulfur atoms intersected by a square with nitrogen atoms arranged 
along the edges (26). Several recent measurements have been made of 
physical properties (27-29); the infrared spectrum of each of the 
trimeric compounds shows a band characteristic of the particular ring 
system (19, 22, 30, 31). 

C. Some Reactions 

Reduction of borazene derivatives yields the corresponding hydro- 
derivatives (32); Grignard reagents have, been used to convert, for 
example, B 3 Cl 3 N3Me3 to B3Et 3 N3Me 3 ('32,33). Substituted borazenes have 
also been prepared directly (3^,35), as by the action of borohydride 
ion on anilinium chloride to produce B 3^^3(06^)3. A polycondensate 
of a borazene has recently been reported (35a), as has the nature of 
the pyrolysis product of the borazene -methanol adduct (35b). The 
'CN) 3 ring of (HCN) 3 is quite reactive, as it is cleaved by hydrazine 
36), primary amines (37), amide ion, and by Friedel-Crafts Catalyst 
38). Cyanuric chloride has been araintiy zed (39) and generally acts 
as an acid chloride; the fluoride is prepared by the action of 

2 ] . 




: ' 



I i 



! ' 







SbF 3 Cl 2 or KS0 2 F on the chloride (40-42). The reaction of the 
chloride and fluoride with various nucleophiles has also been studied 

The azide (43), hydrazide (44), and isothiocyanate (45) of the 
(PN)3 ring have recently been prepared from the chloride. Trimeric 
(NPC1 2 )3 undergoes Grignard and Friedel-Crafts reactions; the methyl 
(46) and phenyl (47) derivatives have also been made directly. 
Various alkylamino derivatives have been prepared, several of which 
have been shown to be isomeric (48-50). The behavior of the chloride 
and fluoride toward several basic reagents has recently been studied 
(50a). Tetrasulfur tetranitride is reduced to S 4 (NH) 4 by stannous 
chloride in ethanol; with acetyl chloride it yields S4N3CI (51) • 
Fluorination yields S 4 N 4 F 4 , while chlorination yields S3N3CI3 (15,52). 
Transition metal derivatives of N 4 S 4 have been made (53), as well as 
several alkyl derivatives (54,55). A compound similar to S 4 (NH) 4 
is heptasulfurimide, S 7 NH (56,57). 

II . Aromaticity 

In all these compounds the Z-N bond distances are individually 
equal, and the trimers are planar. This evidence, coupled with an 
extensive examination of their chemical properties, has been the 
basis from which a discussion of aromaticity in these systems began. 
Borazene is isosteric with benzene, and appears to resemble it re- 
markably in physical properties, while in the P-N and S-N rings there 
is a possibility of p-d pi-bonding and consequent derealization. 
This phenomenon has been explored, using the molecular orbital theory 

III. Quadrupole Resonance 

Cyanuric chloride (60-65) has been studied extensively through 
its quadrupole resonance frequency of absorption, as has (NPCl2)3 
(66,67) and (C1SN) 3 (68). This work has demonstrated the double-bond 
character of the bond between chlorine and a ring atom, and has shown 
that all the chlorines in these molecules are not crystallographically 
equivalent . 

.. f 

■ ' 







. " 




1. G. W. Schaeffer, R. Schaeffer, and H. I. Schlesinger, J. Am. 
Chera. Soc, £3, 1612 (1951). 

2. R. Schaeffer, M. Steindler, L. Hohnstedt, H. S. Smith, Jr., 
L. B. Eddy, and H. I. Schlesinger, ibid., j[6, 3303 (195*0. 

3. W. V. Hough, G. W. Schaeffer, M. Dzurus, and A. C. Stewart, 
ibid., XL, 864 (1955). „ , 

4. C. A. Brown and A. W. Laubengayer, ibid., 77, 3699 (1955). 

5. T. C. Bissot and R. W. Parry, ibid., XL> 34ol (1955). 

6. H. S. Turner and R. J. Warne, Chem. and Ind. (London), 526 (1958) 

7. C. Grundmann and A. Kreutzberger, J. Am. Chem. Soc, ][6, 632 

8. C. Grundmann and A. Kreutzberger, ibid., 76, 5646 ( 1954) . 

9. C. Grundmann and E. Kober, J. Org. Chem., 21_, 641 (I956) . 

10. L. P. Audrieth, R. Steinman, and A. D. F. Toy, Chem. Revs., 32, 
109 (1943). 

11. L. F. Audrieth, J. Chem. Educ. 34, 545 (1957). 

12. L. F. Audrieth, Rec. Chem. Progr., in press. 

13. R. G. Rice, L. W. Daasch, J. R. Holden, and E. J. Kohn, J. Inorg. 
Nucl. Chem., j5, 190 (1958). 

14. F. Seel and J. Langer, Z. anorg. u. allgera. Chem., 29 5 , 316 

15. M. Goehring, Quart. Revs. (London), 10, 437 (1956); M. Goehring, 
Ergebnisse und Probleme der Chemie der Schwefelstickstoffver- 
bindungen, Berlin: Akadamie-Verlag, 1957. 

16. H. Garcia-Fernandez, Bull. soc. chim. France, 265 (1958) . 

17. 0. Westphal, Pharm. Acta Helv., 33, 429 (1958). 

18. K. P. Coffin and S. H. Bauer, J. Phys. Chem. 52, 193 (1955). 

19. J. Gonbeau, E. L. Jahn, A. Kreutzberger, and C. Grundmann, 
ibid., J58, 1078 (1954). 

20. Y. Akimoto, Bull. Chem. Soc. Japan, 2§, 4 (1955). 

21. L. 0. Brockway and W. W. Bright, J. Sm. Chem. Soc, 65, 1551 

22. L. W. Daasch, ibid., £6, 3403 (1954). 

23. A. Wilson and D. F. Carroll, Chem. and Ind. (London), 1558 (1958) 

24. J. A. A. Ketelaar and T. A. DeVries, Rec. trav. chim., 58, 1081 

25. R. L. Sass and J. Donohue, Acta Cryst. 11, 497 (1958) . 

26. D. Clark, J. Chem. Soc, 1615 (1952). 

27. C. W. Rector, G. W. Schaeffer, and J. R. Piatt, J. Chem. Phys., 
17, 460 (1949). 

28. L. B. Eddy, S. H. Smith, Jr., and R. R. Miller, J. Am. Chera. 
Soc, XL> 2105 (1955). 

29. A. R. Humphries and G. R. Nicholson, J. Chera. Soc, 2429 (1957). 

30. W. C. Price, R. D. B. Fraser, T. S. Robinson, and H. C. 
Longuet-Higgins, Discussions Faraday Soc, £, 131 (1950). 

31. R. A. Shaw, Chem. and Ind. (London), 54 (1959). 

32. J. H. Smalley and S. F. SfcafieJ, J. Am. Chem. Soc, 81., 582 


33. G. E. Ryschkewitsch, J. J. Harris, and H. H. Sisler, ibid., 80, 
4515 (1958). 

34. S. J. Groszos and S. F. StafieJ, ibid., 80, 1357 (1958). 

35. H. J. Becher and S. Frick, Z. anorg. u. allgem. Chem., 295* 
83 (1958). 

35a. M. F. Lappert, Proc Chera. Soc, 59 (1959). 

• - 

. • 













35b. D. T. Haworth and L. F. Hohnstedt, J. Am. Chem. Soc, 8l_, 842 

r (1959) • o , 

36. C. Grundmann and A. Kreutzberger, ibid., 7£, 2839 (1957). 

37. C. Grundmann and A. Kreutzberger, ibid., £7, 6559 (1955). 

38. C. Grundmann and A. Kreutzberger, ibid., J7, 44 (1955). 

39. J. T. Thurston, J. R. Dudley, D. W. Kaiser, I. Hechenbleikner, 
F. C. Schaefer, and D. Holm-Hansen, ibid., 73, 298|(195l). 

40. C.J. Grundmann and E. H. Kober, U. S. Patent 2, 845, 421; 
C. A., £3, 1390f (1959). 

41. A. F. Maxwell, J. S. Fry, and L. A. Bigelow, J. Am. Chem. Soc, 
80, 548 (1958). 

42. D. W. Grisley, Jr., E. W. Gluesenkamp, and S. A. Heininger, 

J. Org. Chem., 23, 1802 (1958); F. Seel and K. Ballreich, Chem. 
Ber., 22, 5 44 (1959). ,. 

43. C. J. Grundmann and R. Ratz, Z. Naturforsch., 10b, 116 (1955). 

44. R. J. A. Otto and L. F. Audrieth, J. Am. Chem. Soc, 80, 3575 

45. R. J. A. Otto and L. F. Audrieth, ibid.. 80, 589^ (1958). 

46. H. T. Searle, Proc. Chem. Soc, 7 (1959). 

47. C. P. Haber, D. L. Herring, and E. A. Lawton, J. Am. Chem. Soc, 
80, 2116 (1958). 

48. R. A. Shaw and G. Stratton, Chem. and Ind. (London), 52 (1959). 

49. S. K. Ray and R. A. Shaw, ibid., 53 (1959). 

50. M. Becke-Goehring and K. John, Angew. Chem., JO, 656 (1958). 

51. A. G. MacDiarmid, ibid., £8, 3871 (1956). 

52. 0. Glemser, H. Schroder, and H. Haeseler, Z. anorg. u. allgem. 
Chem., 222, 2 8 (1955). 

53. T. S. Piper, J. Am. Chem. Soc, 80, 30 (1958). 

54. B. D. Stone and M. L. Nielsen, ibid., ££, 1264 (1958). 

55. B. D. Stone and M. L. Nielsen, ibid., in press. 

56. M. Becke-Goehring, LT. Jenne, and E. Fluck, Chem. Ber. 01. 1947 

57. M. Becke-Goehring and R. Schwarz, Z. anorg. u. allgem. Chem., 
296 , 3 (1958). 

58. D. P. Craig/;and N. L. Paddock, Nature, 181 , 1052 (1958). 

59. D. P. Craig, Chemical Society Symposia, Bristol, 1958. P. 343. 

60. H. Negita and S. Saton, J. Chem. Phys., 2£, 602 (1957). 

61. H. Negita, S. Saton, T. Yonezawa, and K. Fukui,';Bull. Chem. Soc. 
Japan, 3£, 721 (1957). 

62. Y. Morino, J. Phys. Soc. Japan, 13, 869 (1958). 

63. M. J. S. pewar and E. A. C. Lucken, J. Chem. Soc, 2653 (1958). 

64. P. J. Bray, S. Moscowitz, H. 0. Hooper, R. G. Barnes, and S. L. 
Segal, J. Chem. Phys., 28, 99 (1958). 

65. F. J. Adrian, ibid., 2£, 1381 (1958). 

66. H. Negita and S. Saton, ibid., 24, 621 (1956). 

67. K. Torizuka, J. Phys. Soc. Japan, 11_, 84 (1956) . 

68. H. Negita and Z. Hirano, Bull. Chem. Soc. Japan, 31, 660 (1958). 
5ta. A. 1. lurg and A. P. Caron, J. Am. Chem. Soc, 81, 836 (1959). 




C. Y. Fujikawa April 7, 1959 

I. Introduction, 

A problem of increasing importance in modern coordination 
chemistry entails the separation and analysis of reaction mixtures of 
robust complexes. Because classical and spectrophotometric methods 
can only be used in special cases, the problem is attacked from a 
different angle, by the adsorption methods. 

Historically, the adaptation of adsorption to analytical and 
preparative procedures is credited to Dr. M. Tswett, a Russian 
botanist, who in a series of three articles published in 1906 (l), 
first called attention to the possibilities of this phenomenon and 
devised a method for its use. Except for a few sporadic applications, 
it lay buried in the literature for 25 years. Its utilization in 
1931, by Kuhn and Lederer (2), for the resolution of complex organic 
mixtures awakened the scientific world to its almost unlimited possi- 
bilities . About five years later, Schwab and his co-workers ( 3) 
applied chromatography to inorganic separations . 

E. Lederer and M. Lederer (4) distinguish between three types of 
chromatography; adsorption chromatography, ion exchange chroma- 
tography and partition chromatography. This distinction is however 
more or less arbitrary. 

II. Adsorption Chromatography. 

Although the columnar adsorption procedure appears to be simple 
and straightforward, its efficiency and its applicability are In- 
fluenced by many interrelated factors . 

Certain substances, such as alumina, magnesia, silica, carbon, 
starch and many others, both organic and Inorganic, have the property 
of attracting to and retaining other substances on the surfaces of 
their particles . In a purely physical type of adsorption, where the 
attraction between the adsorbent and adsorbate is considered to be 
a function of the Van der Waals forces, the solute is usually easily 
eluted and is recovered in its original form. There exists another 
type, which is termed chemi sorption, where upon elution or de- 
sorption, the solute is less easily eluted and may not be recovered 
in the form in which it was first adsorbed. Strain (5) explains the 
mechanism of chromatographic analysis as a dynamic process which de- 
pends on the repeated sorption, desorption and resorption of the 
solute to be separated. 

A. The Column 

The adsorption column can be of almost any size, shape, 
length, or design to suit the problems or whims of the analyst . Both 
pressure and vacuum may be used. Pressure is preferable since vacuum 

I ! 

Lbs ' 


■ ■■ . . 




! - 






■ , . . 






tends to boil off the eluant and disrupt the column* 

B. Adsorbents 

No universal adsorbent has yet been found nor has a perfect 
adsorbent for any given purpose been developed. However, as a result 
of experience with many different solids, the more desirable charac- 
teristics of an adsorbent have now been developed. 

The most satisfactory adsorbents are the alkaline earth oxides, 
hydroxides, and their salts, such as carbonates and sulfates. 
Alumina, carbon, ferric oxide and paper are also commonly used. Talc, 
silica, soda ash and various silicates have been found useful for 
special purposes. Advances have also been made in the utilization of 
organic adsorbents for the resolution of inorganic compounds. 

C . Solvents and Eluants 

There is no sharp demarcation between solvents and eluants . 
Liquids that serve for the adsorption of some compounds may be used 
for the elution of other substances with lower adsorbabilities . 
Choice of a solvent is determined to a large extent by the solubility 
of the materials to be adsorbed and by the activity and solubility of 
the adsorbent. Eluants, as a rule, are chosen with the object of 
obtaining rapid and complete liberation of adsorbed compounds . Some 
of the eluants, arranged in order of increasing polarity, that have 
found extensive use are, pentane, cyclohexane, benzene, chloroform, 
ether, acetone, alcohol, and acetic acid. Eluants of intermediate 
polarity may be obtained by using mixtures of solvents (6) . 

III. Separation of Cobalt (ill) and Chromium (III) Amraines . 

Jensen, Bjerrum and Woldbye (7) have reported the development of 
a useful method for analyzing aqueous mixtures of robust complexes. 
This method, a modification of the adsorption method, is called dis- 
placement analysis . 

The separations were found to be highly dependent on the salt 
applied, bein g i n some cases limited to special kinds of electrolytes. 
Thus Co(NH 3 ) 6 could be separated quantitatively from Cr(NH 3 ) 6 
with 0.01 M KOH, 1 M KC1 as a displacing agent with much better re- 
sults than with basic solutions of NaCl or N(C2H 5 ) 4 C1. On the other 
hand, alk aline NfC aHs ^Cl was necessary for separating mixtures of 
Coen 3 and Cren 3 . 

For quantitative studies the collected samples of a given com- 
ponent are diluted to a known volume in a measuring flask and the 
absorption curve of the solution measured with a recording 0ary 
spectrophotometer. The absorption curve of the initial solution of 
the pure component diluted to the same extent is measured with the 
same instrument. In this way information about the quality of the 
separations is obtained. 

In the analysis of the mixture of cobalt (III) ammines (Hexammine, 
hydroxopentammine and dihydroxotetrammine) the hexammine and pentam- 
mine are first displaced leaving the tetrammine on the column; they 
are then separated on another column. In studying the decomposition 
of chromium(lll) amraines, Jorgensen and Bjerrum (8) have used a .. 









procedure for separating hexaramine, hydroxopentammine, and dihy- 
droxotetrararaine which in principle is similar to that described for 
the corresponding cobalt (III) complexes. 

The chromatographic method is also suitable for the separation 
of cis and trans complexes. An alkaline equilibrium mixture of the 
cis and trans dihydroxobis-ethylenediaraine-cobalt(lll) ions have 
been analyzed. The trans compound was first displaced by means of 
0.1 M NaOH, 1 M NaCl, and the cis compound afterwards by means of the 
stronger displacing agent 0.1 M NaOH, 1 M Li CI . Woldbye (9) has 
determined the cis-trans equilibrium constant for the dihydroxobis- 
e thy lene diamine -chromium (III) ions using a similar procedure. 

The well known series of nitroammine complexes, Co(NH 3 ) 6 
(N0 2 )£~ , where xrO, 1, 2, 3 and 4, have been examined. The con- 
ditions under which they can be separated in cases where two or three 
of the complexes are present simultaneously, have been established. 


1. M. Tswett, Ber. 24, 234-44, 316, 384-92 (1906). 

2. R. Kuhn, and E. Lederer, Ber. 64, 1349 (1931). 

3. G. M. Schwab, and K. Jockers, Angew, Chem. 50, 546-53 (1937). 

4. E. Lederer and M. Lederer, "Chromatography", Elsevier Publishing 
Company, New York, 2nd Ed., 1957. 

5. H. H. Strain, "Chromatographic Adsorption Analysis", Interscience 
Publishers, Inc., New York, Ch. 3, 1945. 

6. 0. C. Smith, "Inorganic Chromatography", D. Van Nostrand Co., 
Inc., New York, Ch. 8, 1953. 

7. A. Jensen, J. Bjerrum, and F. Woldbye, Acta Chem. Scand., 12, 
1202 (1958). — 

8. E. Jorgensen and J. Bjerrum, Acta Chem. Scand., 12, 1047 (1958). 

9. F. Woldbye, Acta Chem. Scand., 12, 1079 (1958). 


J. S. Oh April 21, 1959 

I . Introduction 

It is well known that a solution of ferric salt becomes colorless 
on the addition of orthophosphoric acid. This reaction has been wide- 
ly utilized in analytical chemistry. The decolorizing effect has been 
attributed to the formation of a colorless complex or of complexes of 
ferric iron with phosphoric acid by Weinland and Ensgraber (l). 
Carter and Clews (2) showed that phosphoric acid lowers the oxidation 
potential of Fe - Fe system. Bonner and Romeyn (3) observed 
that the reversible reaction Fe + 1/2 I 2 (aq.) J Fe +1" goes to 
completion in the presence of phosphoric acid. All the complexes re- 
ported do exist, under different conditions. Ferric phosphate com- 
plexes are derived from the [Fe(H20) 6 ] ion by the substitution of 
phosphate ions, and possibly of other anions present in the solution, 
for the water molecules in the complex. Van Wazer and Callis (4) has 
recently reviewed metal complexing by phosphates . 

II. Physical Methods 

There are many methods which can be applied to the investigation 
of complex formation ( 5) . 

A. Spectrophotometry Methods. 

The spectrophotometric method is especially well suited to the 
study of complexes not sufficiently stable to permit their isolation 
from solution. Work of this type has been done by Job, who developed 
the Method of Continuous Variations, and extended by Vosburgh and 
Cooper (6), particularly to deal with the formation of more than one 

Lanford and Kiehl (7) determined the composition of the iron- 
phosphate complex by applying Job's method in ferric nitrate-;-, 
phosphoric acid mixtures employing SCN~ solution for the coloriraetric 
determination of ferric ion. According to the data, the composition 
corresponds to a 1:1 complex, i.e., Fe(HP0 4 ) . Its dissociation 
constant, 4.44 x 10" 10 , has been evaluated at an ionic strength of 
O.665 and 30°. These data cannot be considered reliable since the 
formation of different iron thiocyanate complexes was not taken into 
account . 

Banerjee (8) used several different physico-chemical methods; 
the composition which he determined for the complex agreed with 
Lanford 's result. But he detected, in addition to Fe(HP0 4 ) , another 
complex having the formula Fe(HP0 4 ) 2 ~. 

Pilipenko and Ivashchenko (9) studied the complexes by using a 
colorimetric method in the ultraviolet region, and found the complex 
with the ratio Fe : P0 4 " 3 of 2:1. 




Ram, Bose and Kumar (10) concluded that the complexes are proba- 
bly [FeCl(H 2 P0 4 ) 3 ] ~ in the presence of chlorides and [FeS0 4 (H 2 P0 4 ) 2 ] " 
in the presence of sulfates. When monochromatic light passes through 
a transparent, colored medium, part of it is absorbed, part is re- 
flected and the rest is transmitted. Transmission (T) is related to 
the concentration of coloring substance according to the equation: 
T = Re* , kcl = D. Optical density was calculated from the re- 
spective values of T using the above equation and was plotted against 
the molar concentration of phospheric acid present in the different 
solutions . 

B. Electrometric Methods. 

Among the important quantitative data obtainable by polarographic 
means are dissociation constants of complexes, coordination numbers 
of metal ions, and the degree of stabilization of various oxidation 
states. Polarographic studies were carried out by using the support- 
ing electrolytes because of the reduction of ferric ions by mercury 
by Ram, Bose and Kumar (11). It has been shown that one of the com- 
plexes formed in the system FeP0 4 - H 3 P0 4 - NaH 2 P0 4 is [Fe(H 2 P0 4 ) 4 ] ~ 
and the_dissociation constant of the complex containing four 
(H 2 P0 4 )~ groups is the order of 10~ 9 . 

Dede (12) used a conductivity method for the determination of 
phosphate complexes in solution. A sharp increase in conductivity 
upon combination of H3P0 4 and FeCl3 was explained by the formation of 
a very strong acid H 3 [FeP0 4 Cl 3 ] • 

Ricca and Meduri (13) have obtained a maximum at a 1:1 mole 
ratio from conductometric measurement of FeCl3 - H 3 P0 4 mixtures, and 
have formulated the compound as H 3 [FeP0 4 Cl 3 ] in agreement with Dede's 
result . 

Jensen (14) also discussed the decoloration of a solution of 
FeCl 3 with H 3 P0 4 and disagreed with the conclusions of others. From 
the results of his own experiments entailing solubility and conduc- 
tivity measurements, he concluded that the complex is a pure phosphate 
complex and not a mixed chlorophosphate complex. 

Potentiometric titrations commonly used in determining the end 
point of precipitation and oxidation -re duct ion reactions are adapted 
for studying the complex-forming reactions. 

Ram and his coworkers carried out titrations with solutions of 
different ferric ion concentrations. From the values of the e.m.f. 
(E) obtained for different volumes of the standard phosphoric acid 
added (V), values of dE/dV were calculated and plotted against corre- 
sponding values of V. From the amounts of H 3 P0 4 added at these peaks, 
the corresponding ratios of H 3 P0 4 /Fe were calculated. 

Conductometric measurements of dilute solutions of FeCl 3 to which 
varying amounts of concentrated ! H 3 P0 4 i- have been added suggest 
formation of 1:1 and 1:2 complexes which have been formulated as 
[Fe(HP0 4 )] and [Fe(HP0 4 ) 2 ] ~. Transport experiments indicate that 
cationic and anionic complexes are present since iron migrates to both 
cathode and anode. 



C . Other Methods . 

Banerjee also studied this problem by applying the thermometric 
titration method proposed by Haider (15). He obtained three breaks 
in the thermometric titration curve corresponding to FeCl 3 :H 3 P04 
ratios of 1:1, 2:3 and 1:2. The break corresponding to the 2:3 ratio 
was considered to represent a mixture of 1:1 and 1:2 and this might 
not represent a new compound. 

Salmon and coworker (16) studied the complexes by using ion ex- 
change resins. They concluded that the complex ion contains at least 
three phosphate groups and can be represented by either of the form- 
ulas, [Fe(HP0 4 )33~ 3 or [Fe( PO^a] " 6 . Evidence for the presence of 
ions with a lesser number of phosphate groups was stated to be not 
very convincing. 

Ill . Conclusion 

A number of complex ions are formed but their nature has not 
been definitely established. The concept of the formation of soluble 
complexes — whether by strong ion association or covalent bonding — 
has been used to the present day for interpreting various chemical 
anomalies found in phosphate solutions. The higher values of pK, 
obtained for the iron complexes may involve iron-oxygen bonds ex- 
hibiting appreciable covalent character. 


1. R. F. Weinland and F. Ensgraber, Z. anorg. Chem., 84, 340 (1914). 

2. S. R. Carter and F. H. Clews, J. Chem. Soc . , 125, 1880 (1924). 

3. W. D. Bonner and H. Romeyn, Ind. Eng. Chem., Analy. Ed., 3, 85 

4. J. R. van Wazer and C. F. Callis, Chem. Revs., 58, 1011 (1958). 

5. J. C. Bailar, "Chemistry of the Coordination Compounds", pp. 563- 
624, Reinhold Pub. Co., New York (1956). 

6. W. C. Vosburgh and G. R. Cooper, J. Am. Chem. Soc, 63, 437 (194l) 

7. 0. E. Lanford and S. J. Kiehl, J. Am. Chem. Soc, 64, 291 (1942). 

8. S. Banerjee, J. Ind. Chem. Soc, 2£, 417 (1950). 

9- A. T. Pilipenko and L. N. Ivashchenko, J. Gen. Chem., U.S.S.R, 26, 
751 (1956). — 

10. A. Ram. A. K. Bose and S. Kumar, J. Sci. Indust. Res. India, 15B, 
78 (1956). — ' 

11. A. Ram, A. K. Bose and S. Kumar, ibid., 13B , 217 (1954). 

12. L. Dede, Z. allg. anorg. Chem., 125, 28 (1922). 

13. B. Ricca and Meduri, Gazzetta, 6J7~235 (1934), C.A., 28, 5001(1934) 

14. K. A. Jensen, Z. anorg. Chem., 221 , 1 (1934). 

15. B. C. Haider, J. Ind. Chem. Soc, 23, 147 (1946). 
B. C. Haider, ibid., 23, 153 (1946)7 

16. J. E. Salmon, J. Chem. Soc, 74, 2316 (1952). 
J. E. Salmon, ibid., 1953 , 26W. 

R. F. Jamson and J. E. Salmon, ibid., 1954, 28. 





C9 1 







R. E. Eibeck 


April 28, 1959 

Progress in the science of rocketry during the past twenty years 
has been so rapid that laymen have had a difficult task in trying to 
keep up with developments. Even though, at the present, almost every 
so-called news publication abounds with references to rocketry it is 
doubtful whether sufficient basic knowledge of rocketry is generally 
available to appreciate and understand the new developments in the 



Before proceeding to a discussion of solid propellants it may be 
helpful to present some basic information on rocketry. 

The chief purpose of the rocket engine is to deliver thrust or 
force in a given direction for a given period of time. The thrust 
is obtained from the ejection of microscopic particles generated 
completely from substances carried within the system itself. This 
method of operation depends upon the physical laws of action and 
reaction and may be expressed by: 

F = m a 

in which the force or thrust F is equivalent to the product of a mass, 
m, and the acceleration of that mass, a. In a rocket the burning 
propel lant gives gases or solids at high temperatures and pressures. 
The gases are accelerated from the engine ( m x a) to produce thrust. 

The purpose of a rocket engine is to produce an impulse, i.e ., 
the integral of thrust for a given period of time, t. When the thrust 
Is constant the impulse, I is given by: 

1 = F t 

A useful method of evaluating performance is the propel lant specific 
impulse, I sp , which is defined as: 

sp % 

where W is the propellant weight. Specific impulse is usually 
expressed in seconds. 

Further it may be shown that 

V awT /M 
ex c 

where V ex is the actual exhaust velocity, T , the combustion chamber 
temperature, and M, is the average molecular v/eight of the combustion 

The relationships presented above serve to define two desirable 
properties for rocket propellants: 5 







■ • 


• . ■ 


. j 

. . . s J 



. t 









(1) Small negative or preferably positive heat of formation. 

(2) Low molecular weight and large negative heat of forma- 
tion for the reaction products, 


A. Monopropel lants (Single Base) — 

A monopropel lant is a single chemical compound which 
is stable at normal temperature and pressure but v/hich is sufficiently 
reactive at elevated temperatures and pressures to convert the heat 
Df decomposition of the compound into jet kinetic energy without 
reaction with external compounds. This type of propellant is repre- 
sented by cellulose nitrate. 

B. Homogeneous Propel lants (Double Base) — 

A homogeneous propellant is a propellant in which the 
oxidizing and reducing agents (or a mixture of monopropel lant materials) 
Dccur as a single, or colloidal phase. An example of this type of 
propellant is cordite in which cellulose nitrate is colloided with 
another monopropel lant such as glycerol trinitrate or DEGN. 

C. Heterogeneous Propellants (Composite) 

Such propellants are composed of a mixture of oxidant 
and reductant, each of which is present as a separate phase. Common 
Dlack powder is an example of this type of propellant. 

For many years homogeneous propellants were used almost exclu- 
sively; it is only quite recently that the more powerful composite 
propellants have come to the fore. 


After a solid propellant is compounded it must be given a 
specific shape for use in a rocket. There are many possible shapes 
for solid propellant charges, each having its advantages and dis- 
advantages. For maximum impulse it would be desirable to fill the 
entire rocket chamber with propellant and burn the propellant from 
the end, like a cigarette. This would be fine if a propellant 
imposition could be found which would burn at a rate of 30 - 100 
Ln./sec, however, current propellants burn at 2 in. /sec. or less. 
In order to obtain burning times of the required duration, it is 
lecessary to restrict the total burning distance. This means that 
Durning must occur radially in a cylindrical cnarge rather than 

From a practical viewpoint the many different charge shapes may 
be divided into three classes, each characterized by the method by 
which it burns. They are: (l) internal - external burning charges, 
(2) external burning charges, and (3) internal burning charges* 
lach of these charge shapes possesses its own specific burning 
character i sties. 


( l) Arendale, V/.F. , Fuel Binder Requirement for Composite Propellants , 
Ind. Eng. Chem. Ig7~725 1 1955"H 







• - . 


(2) Butz, J.S., Solid Fuels Contend for Long Range Bal li stic Role , 
Aviation Week 67, No. 11+, SO U9FT7. 

(3) Missiles and Rockets . 2, (August 195?) Entire issue. 

(I4.) Penner, S.S., Chemical Propel lants , J. Chem. Ed. 29, 37 (1952). 

(5) Penner, S.S., Chemistry Problems l"n Jet Propulsion , Ch. 7* 
Pergamon Press, N. Y. 1 1 95>TI 

(6) Ritchey, H.W., Solid Propel lants Vie With Liquids for Rockets , 
Chem. and Eng. News 35, No. hS, 7B^H2~TT957TT^ 

(7) Thompson, R.L., Rocket Propel lants , Chem. and Eng. News 36, No. 
25, 62-7 (1958). "~~ 

(8) Warren, F.A., Rocket Propel lants , Reinhold Publishing Corp., 
N.Y. 1958. 

(9) Wimpress, R.N. , Internal Ballistics of Solid Fuel Rockets , McGraw 
Hill Book Co., Inc., N.Y. 1950. 

10) Zaehringer, A.J., Solid Propellant Rockets , Amer. Rocket Co., 
Wyandotte, Michigan, 195BT 



N. J. Rose M ^ 5, 1959 


The desire for bigger and better rockets has greatly stimulated 
a variety of research and development activities, including studies 
of the fundamentals of combustion, 10 development of high energy fuels, 
and the development of refractory coatings for rocket engines. 11 
Moreover, the rocket industry has had a profound effect in many non- 
scientific areas as is suggested by the following statement taken 
from a recent issue of C.and E.News : 

The U. S., reacting from a swift kick in the 
scientific pants from Russia on October k 9 
1957* grimly started out to make 195° a year 
of catching up. More than national pride 
as the world 1 s scientific leader was involved; 
at stake was survival itself. 6 


Thus, the political and strategic implications of the rocket race 
are currently subjects of great interest. * The economy of the nation 
has also felt the impact of the rocket industry. The national missle 
program consumes 20% of the nation; s oxygen output, has greatly 
stimulated the perchlorate market, 12 and in general Dours approximately 
2,500 millions of dollars into the economy annually. 

In addition to international competition there is a good deal 
of rivalry within the rocket industry itself, concerning the relative 
merits of liquid and solid propellant units. According to John P. 
Longwell, Esso Research and Engineering, liquid propellants will 
always give a slightly more favorable performance [I ) but the 
reliability of solid fuels will make them more desirable for military 
purposes. 1 5 

Liquid propellants may be classified as monopropel lants or 
bipropellants. A monopropel lant does not require the addition of 
another ingredient to bring about the release of its thermochemical 
enerqv. The bipropellant has two compounds and may be either hyper- 
golic (self-igniting) or nonhypergolic. 3 A consideration of the 
desirable properties of liquid propellants provides a convenient for- 
mat for their discussion. 


A. Propellants should have small negative or preferably posi- 
tive standard heats of formation and should form reaction products 
with large negative heats of formation and with low molecular 
weight. i'3»7 The performance of a propellant system is measured as 
specific thrust, I SD , which is the thrust per unit weight rate of 
flow of the propellant. The specific thrust is a function of the 
square root of the term T c /M where T c is the combustion temperature 
and M is the average molecular weight of combustion products. 13 ' 1 * 

■ ■ ; • 


• '• : 







lb. /sec. 

liquid hydrogen 



ethanol 100% 



ani line 


High heats of formation would tend to give a high T c and low molecular 
weight products would make the T c /M ratio large. Both of these 
criteria are satisfied when the propellants of a system are either the 
elements of the first rows of the periodic chart or compounds com^ 
posed of these elements. Table I contains some representative per- 
formance values for particular propellants, 


Oxidant Fuel Mixture Ratio 0/F Specific Thrust 

Liquid oxygen 
Liquid oxygen 

Attaining maximum thrust necessitates optimum combustion 
conditions. This means that a smooth feeding system and efficient- 
injection system are required. 

1« Feed Systems 

a. Simple pressure feed systems utilise the pressure 
from the boiling oxidant (e.g., liquid oxygen), from a high pressure 
third tank, from a small solid propellant cartridge, or from spon- 
taneous ignition of part of the main fuels to drive the fuel into the 
combustion chamber, 3 * 4 

b. The turbo-pump system uses the gases from a small 
generator to drive a turbine which in turn drives the fuel pumps. 
The gas generator may use solid propellants, a liquid propellant 
different from that used in the main rocket, or it may use one of 
the components of the main propellants (e.g., the gas obtained by 
passing H 2 2 from main oxidizer tank over a surface catalyst), 

2, Injectors 

a. A shower head injector provides a fine spray and 
therefore rapid evaporation rates. It is useful for nonhyper golic 
propellants where the combustion rate is probably controlled by a 
gas-phase chemical reaction. Mixing of the liquids would have little 
effect on the combustion rate. 1 ' 3 ' 6 

b. A liquid-liquid impinger is commonly used for 
hypergolic bipropellant systems because any operation which facili- 
tates liquid-liquid collisions and liquid-phase chemical reaction 
rates will allow maximum combustion rates, 

c. Splash plates are often used when the reaction is 
surface catalyzed. 1 




' ' 



- . 


•i . ■■> ■ 



' ' 


. ■ - . ! 


Great care is taken to design the feeding and injection systems 
so that high-frequency pressure oscillations do not develop during 
combustion. 3 Frequently these oscillations can become severe enough 
to destroy the rocket, 

B. Propel lants should have high densities so that dead weight 
of storage tanks and the increased size of the rocket as a whole 
may then be avoided. 

C. At least one of the propel lant components should have a 
relatively high boiling point and high specific heat. 1 The high 
temperature produced by the operation of a rocket requires that the 
chamber walls be protected in some way. Several cooling systems have 
been successfully used. 4 ' 

1. Concentric feed. In this system, the fuel is injected 
so that it forms a concentric cylinder around the injected oxidant 
and therefore protects the chamber wall to some extent. 

2. Regenerative cooling. One of the propel lant components 
flows through the~annular passage between the chamber and the rocket 
wall before it is introduced into the chamber. In addition to cooling 
the chamber wall, this process increases the efficiency of the motor 
by reducing heat loss. This system can be modified simply by coiling 
the fuel lines around the chamber and nozzle. 

3. Film cooling. Here, a portion of the fuel is injected 
through a series of orifices into the chamber to form a cooling film 
on the inner surface. 

D. Propellants should be non-toxic, should not decompose or 
change chemically during storage, and should be readily available 
and relatively cheap. Only the last of these considerations requires 
comment. The price of liquid fluorine would drop from 5*00 dollars 
to 1.00 dollar per pound if there were a demand for daily multiton 
shipments. 7 


The ignition and thrust of a rocket are related by a seriesof 
empirical correlations for which there is no satisfactory theoretical 
model. Thus, we see that the primary performance characteristic, 
the specific thrust, depends not only upon the chemical and thermo- 
dynamic properties of the propellant but also on the design of the 
various functioning units. These empirical relationships have 
provided a satisfactory basis for the development of successful 


1. Penner, S. S., Chemistry Problems in Jet Propulsion , Pergamon 

Press, New York, 1957* 

2. Caidin, M. , Countdown for Tomorrow , £. P. Dutton and Co., Inc., 

New York, 1958. 





. ■ 







3, Warren, F, A., Rocket Propel lants , Reinhold Publishing Corp., 
New York, 19^BT 

ij.. Burgess, E., Rocket Propulsion , Chapman c.nd Hall Ltd., London, 

5. Humphries, J., Rockets and Guided Missi les , The MacMillan Co., 

New York, 19^57^ 

6. Guided Missiles Fundamentals , Dept. of the Air Force, AF 

Manual 52-31- 

7. C. and E. News, 35, No. 21, l8 (1957). 

8. C. and E. News, £1, No. 1, 33 (1959). 

9. C. and E. News, 37, N . k, 78 (1959). 

10. Friedman, R., and Grover, J. H., Ind. and Eng. Chem., Ll9, 

IklO (1957). 

11. C. and E, News, 36* No. 50, SO (1958). 

12. C. and E. News, 36, No. 51, 11 (1958). 

13. Clark, G. L., The Encyclopedia of Chem . ( Supplement ) , Reinhold 

Publishing Corp., New York 1958. 

ll\.. Eibeck, R., Seminar,. University of Illinois, 1959. 
15. C. and E. News, 37, No. 13, 22 (1959). 






R. E. Sievers May 12, 1959 

During the past decade infrared spectroscopy has become one 
of the most important tools in the hands of the coordination 
chemist. Organic ligands quite often contain functional groups 
which absorb light in the infrared region. Chelation of such 
ligands produces marked changes in frequency and intensity of 
certain absorption bands. It is sometimes possible to relate 
these changes to some characteristic of the resulting metal 
chelates. Of particular interest are the studies made on chelates 
of ethylenediaminetetra-acetic acid and on acetylacetone and its 

I. Chelates of Ethylenediaminetetra-acetic Acid (EDTA) 

EDTA contains two nitrogen and four oxygen donor 
atoms, making it potentially a sexidentate ligand. Stable 
complexes are formed however in which one or more of the six 
donor atoms are not attached to the metal ion. The presence of an 
uncomplexed carboxyl group may be detected In the infrared 
spectra. Busch and B a i lar 1 prepared cobalt complexes containing 
EDTA which in one case was sexidentate and in two other cases 
was pentadentate. The infrared spectra of the complexes 
containing pentadentate EDTA show two carbonyl bands. One 
band, at 1723 cm"* 1 is attributed to a free carboxyl group. 
A stronger band, appearing at 1628 cm*" 1 , is attributed to the 
presence of three coordinated carboxyl groups. In a compound 
containing a sexidentate EDTA group, only one absorption band 
is observed at 1628 cm" 1 . This indicates that all four carboxyl 
groups are equivalent and coordinated. 

Shifts to lower frequencies upon complex formation have 
been explained in terms of the partial ionic character of the 
coordinate link 1 " 7 . If a carboxyl group is linked to some group 2 

— c— o — z 

the carboxylate resonance will increase as the ionic character 
of the O-Z link increases 2 . This results in a greater local- 
ization of electrons in the carbonyl group, imparting single 
bond character to the group and thus lowering the frequency. 

An infrared study of Nd and Y complexes of EDTA led to the 
conclusion that the EDTA group is pentadentate and that a 
water molecule occupies the sixth coordination position 3 . 
Differential thermal analysis further supported this conclusion. 
Kirschner 4 used infrared spectra as evidence for a coordination 
number of six for the Cu(ll) complex of EDTA, Tetradentate 
and bidentate complexes of EDTA with Pd(ll) and Pt(ll) have 
also been studied 5 . 



Sawyer and Paulsen have attempted to correlate stability 
of alkaline earth chelates wi th shifts in frequency. They also 
observed that the bonding in these complexes is primarily ionic. 
This was similarly found to be the case for complexes of Mn(ll), 
Co(Il), Nl(Il), Zn(ll), Cd(ll), Hg(ll), and Pb(ll) 7 . 

II. Chelates of (3-diketones 

A cyclic, symmetrical, resonance-stabilized structure 
is now generally accepted for metal acetylacetonates 6 . 

Ri R 2 

C ■ C 

I: :l 


The infrared spectra of these compounds is very rich in the 
region 1700 to 280 cm" 1 and assignment of the absorption bands 
is difficult. In the sodium chloride region, bands arising from 
C— O and C-^C stretching modes are of prime interest. 9 in the 
cesium bromide region, bands attributed to M-^0 stretching have 
been reported. 10 

Infrared spectra of chelates of acetylacetone with beryllium, 
magnesium, chromium, iron, cobalt, copper, zinc, thorium, uranium, 
nickel, aluminum, scandium, americium, samarium, boron, titanium, 
zirconiua, cesium, potassium, sodium, lithium, thallium, silver, 
calcium, palladium, and lanthanum have been studied. 8 " 14 

The effect of varying groups Ri and R 2 has been studied in 
Cu and Ni chelates of derivatives of acetylacetone. 9 -i3 

The intense activity in this area may be attributed to attempts 
to relate the stability of these complexes to observed frequency 

Bellamy and Branch found a straight line relationship 
when they plotted the carbonyl frequencies versus the stabilities 
of bivalent metal complexes of salicylaldehyde. They reasoned 
that it should therefore be possible to utilize infrared measure- 
ments to determine stabilities. However when the spectra of copper 
chelates of various p-diketones 1 ,13 were compared to their 
stability, no regular relationship was found. 

Holtzclaw and Collman examined the spectra of acetylacetone 

chelates of several divalent metals and found a qualitative 
relationship between the strength of the M-^-0 bond and the frequency 
of the perturbed carbonyl absorptions. The frequency decreases 
as the stability increases for the series: Mg(ll), Cd(ll), Mn(ll), 
Co(ll), Zn(ll), Ni(ll), Cu(ll). The nickel complex is the only 
exception. It absorbs at a higher frequency than this relation- 
ship would predict. 



West and Riley studied acety lacetonates of univalent, 
divalent, and trivalent metals and attempted to relate the carbonyl 
band frequency to respectively the first, second, and third 
ionization potentials. They were successful for divalent and 
trivalent metals, but the relationship broke down for the 
univalent case. 

i o 
Most recently interest has been shifted to the M-i-0 

frequencies in the potassium bromide and cesium bromide regions. 
Assignment of bands in these regions is very uncertain but pre- 
liminary work indicates that some correlation of complex stabil- 
ities to M— -O stretching frequencies may be possible. Bands 
ranging from Lj.20 to lj.80 cm" 1 have been attributed to metal-oxygen 
stretching modes. 


1. D, H. Busch and J. C. Bailar, Jr., J. Amer. Chem. Soc , 

25, ksik (1953). 

2. M. l. Morris and D. H, Busch, J. Amer. Chem. Soc ,, 76, 

5178 (1956). 

3. T. Moeller, F. A. J. Moss, R. H # Marshall, 77, 3182 (1955). 
]+. S. Kirschner 78, J. Amer Chem. Soc , JQ, 2 372 (1956). 

5« D. H. Busch and J"^ C. Bailar, Jr., J. Amer. Chem. Soc , 

28, 716 (1956). 

6. D. T. Sawyer and P. J. Paulsen, J. Amer. Chem. Soc , 80, 

1597 (1958). 

7» D. T. Sawyer and P. J. Paulsen, J. Amer. Chem. Soc , 8l, 

816 (1959). "~ 

8. J. Lecompte, Disc Farad. Sop , 9, 125 (1950); C. Duval, 

R. Freymann and J. Lecompte, Bull. Soc Chim. France , 
19, 106 (1952). 

9. R. P. Dryden and A. Winston, J. p hys. Chem. , 62, 635 (l958). 

10. K. Nakamoto, P. J. McCarthy and A. E. Martell, Nature , 

183 , 14-59 (1959), 

11. R. L. Belford, A. E. Martell and M. Calvin, J. Inorg. N uc l « 
Chem ., 2, 11 (1956). 

L. J. Bellamy and R. F. Branch, J. Chem. Soc , 195 U- » U491« 
H. F # Holtzclaw, Jr. and J. R. Cpllman, J. Amer. Chem. Soc ., 
79, 3318 (1957). 
111.. R. West and R. Riley, J. 1 norq, Nucl. Chem ., 5, 295 (l958).